You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@guacamole.apache.org by Leo Nikolaev <ni...@kasatkina.org> on 2021/08/02 20:59:14 UTC
Re: guacamole broken on Safari
Okay, I’ve done some triaging by building and checking manually some commits.
Commit 037ed21 is definetely okay, and it even has the logout support, so fits perfectly.
There is a bunch of commits at 8th of July then, and the last of them, dd86130 is completely broken.
Commit cc239b6e has the bug with Safari.
So I think bug was introduced somewhere at 8th of July, will try to test more commits.
Cheers,
Leo
> On 30 Jul 2021, at 04:00, Mike Jumper <mi...@glyptodon.com> wrote:
>
> On Wed, Jul 28, 2021 at 3:36 AM Leo Nikolaev <ni...@kasatkina.org> wrote:
> If comfortable, I can record a screencast or do a Zoom demo of what am I doing.
>
> I’m absolutely sure about latest master - I mount it inside Docker-container, and when I remove -v parameter it all works perfectly on 1.3.0.
>
> docker run --name guacamole \
> --link guacd \
> --link mysql \
> -v $(pwd)/guacamole:/etc/guacamole \
> -v $(pwd)/guacamole-client/guacamole/target/guacamole-1.3.0.war:/opt/guacamole/guacamole.war \
> -e MYSQL_DATABASE=guacamole \
> -e MYSQL_USER=root \
> -e MYSQL_PASSWORD=guacamole \
> -e GUACAMOLE_HOME=/etc/guacamole \
> -d -p 8080:8080 \
> guacamole/guacamole
>
> Building it from repo, yes.
>
> Manually layering an externally-built, newer .war file over the .war within the Docker image from a previous release is a bit funky. To ensure things are sane, can you build the Docker image directly from the same source and get rid of the volume mount? The Dockerfile for the image is part of the guacamole-client repository, so you can just do a "docker build" from the root of the source that you cloned from git to produce an updated image.
>
> Thanks,
>
> Michael Jumper
> CEO, Lead Developer
> Glyptodon Inc.
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
For additional commands, e-mail: user-help@guacamole.apache.org
Re: guacamole broken on Safari
Posted by Mike Jumper <mi...@glyptodon.com>.
This should now be fixed via
https://github.com/apache/guacamole-client/pull/639. The issue was that an
old CSS class, "tiled-client-list", was replaced with a corresponding
directive, "guac-tiled-clients" via commit c990043, but the style rule that
referenced "tiled-client-list" was not updated accordingly.
Michael Jumper
CEO, Lead Developer
Glyptodon Inc <https://glyp.to/>.
On Mon, Aug 16, 2021 at 8:46 PM Mike Jumper <mi...@glyptodon.com>
wrote:
> OK - after installing a copy of OS X "Catalina" within VirtualBox, I'm now
> able to reproduce this on the included older version of Safari. I'll start
> digging into what CSS changes are needed to un-break older Safaris.
>
> Michael Jumper
> CEO, Lead Developer
> Glyptodon Inc <https://glyp.to/>.
>
>
> On Fri, Aug 13, 2021 at 6:03 AM Leo Nikolaev <ni...@kasatkina.org>
> wrote:
>
>> Okay, I think this could be closed.
>>
>> I’ve done the tests, only the old versions of Safari are affected: 14.0
>> and below. As long as we have a relatively small share of old Safafi
>> browsers.
>>
>> Btw, I’ve tracked down the bug in current master, it all comes to wrong
>> heigth, which is not firing from somewhere deep inside tiles.
>>
>> Mike, thanks for your help :)
>>
>> Cheers,
>> Leo
>>
>> > On 13 Aug 2021, at 00:42, Mike Jumper <mi...@glyptodon.com>
>> wrote:
>> >
>> > On Wed, Aug 11, 2021 at 2:52 PM Leo Nikolaev <ni...@kasatkina.org>
>> wrote:
>> > You mean I should not see this issue on latest master? Is there a
>> different code there?
>> >
>> > The code on master should be the same code that you bisected - it just
>> happens that code from the specific commit that you found through the
>> bisect was already effectively undone by a later commit that would also
>> have been in the history when you did the bisect.
>> >
>> > All this means is that the result of the bisect is not as
>> straightforward as would be ideal.
>> >
>> > - Mike
>> >
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
>> For additional commands, e-mail: user-help@guacamole.apache.org
>>
>>
Re: guacamole broken on Safari
Posted by Mike Jumper <mi...@glyptodon.com>.
OK - after installing a copy of OS X "Catalina" within VirtualBox, I'm now
able to reproduce this on the included older version of Safari. I'll start
digging into what CSS changes are needed to un-break older Safaris.
Michael Jumper
CEO, Lead Developer
Glyptodon Inc <https://glyp.to/>.
On Fri, Aug 13, 2021 at 6:03 AM Leo Nikolaev <ni...@kasatkina.org> wrote:
> Okay, I think this could be closed.
>
> I’ve done the tests, only the old versions of Safari are affected: 14.0
> and below. As long as we have a relatively small share of old Safafi
> browsers.
>
> Btw, I’ve tracked down the bug in current master, it all comes to wrong
> heigth, which is not firing from somewhere deep inside tiles.
>
> Mike, thanks for your help :)
>
> Cheers,
> Leo
>
> > On 13 Aug 2021, at 00:42, Mike Jumper <mi...@glyptodon.com> wrote:
> >
> > On Wed, Aug 11, 2021 at 2:52 PM Leo Nikolaev <ni...@kasatkina.org>
> wrote:
> > You mean I should not see this issue on latest master? Is there a
> different code there?
> >
> > The code on master should be the same code that you bisected - it just
> happens that code from the specific commit that you found through the
> bisect was already effectively undone by a later commit that would also
> have been in the history when you did the bisect.
> >
> > All this means is that the result of the bisect is not as
> straightforward as would be ideal.
> >
> > - Mike
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> For additional commands, e-mail: user-help@guacamole.apache.org
>
>
Re: guacamole broken on Safari
Posted by Leo Nikolaev <ni...@kasatkina.org>.
Okay, I think this could be closed.
I’ve done the tests, only the old versions of Safari are affected: 14.0 and below. As long as we have a relatively small share of old Safafi browsers.
Btw, I’ve tracked down the bug in current master, it all comes to wrong heigth, which is not firing from somewhere deep inside tiles.
Mike, thanks for your help :)
Cheers,
Leo
> On 13 Aug 2021, at 00:42, Mike Jumper <mi...@glyptodon.com> wrote:
>
> On Wed, Aug 11, 2021 at 2:52 PM Leo Nikolaev <ni...@kasatkina.org> wrote:
> You mean I should not see this issue on latest master? Is there a different code there?
>
> The code on master should be the same code that you bisected - it just happens that code from the specific commit that you found through the bisect was already effectively undone by a later commit that would also have been in the history when you did the bisect.
>
> All this means is that the result of the bisect is not as straightforward as would be ideal.
>
> - Mike
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
For additional commands, e-mail: user-help@guacamole.apache.org
Re: guacamole broken on Safari
Posted by Mike Jumper <mi...@glyptodon.com>.
On Wed, Aug 11, 2021 at 2:52 PM Leo Nikolaev <ni...@kasatkina.org> wrote:
> You mean I should not see this issue on latest master? Is there a
> different code there?
>
The code on master should be the same code that you bisected - it just
happens that code from the specific commit that you found through the
bisect was already effectively undone by a later commit that would also
have been in the history when you did the bisect.
All this means is that the result of the bisect is not as straightforward
as would be ideal.
- Mike
Re: guacamole broken on Safari
Posted by Leo Nikolaev <ni...@kasatkina.org>.
You mean I should not see this issue on latest master? Is there a different code there?
Cheers,
Leo
> On 12 Aug 2021, at 00:46, Mike Jumper <mi...@glyptodon.com> wrote:
>
> Nope, not as of yet. Based on what you're seeing, it sounds like a CSS issue, yes, but not an issue with grid layout. The code that you're running across as the cause was only an intermediate attempt during development and actually isn't present any longer. The layout was migrated to flexbox prior to merge.
>
> That said, I would expect something CSS-related to be consistent and abundantly reproducible. So far, I can't reproduce it at all in the only copy of Safari I (currently) have access to.
>
>
> On Wed, Aug 11, 2021 at 2:31 PM Leo Nikolaev <ni...@kasatkina.org> wrote:
> Mike, any update on this? Looks like a CSS bug in tile grid :)
>
> Cheers,
> Leo
>
> > On 6 Aug 2021, at 10:42, Leo Nikolaev <ni...@kasatkina.org> wrote:
> >
> > Okay, I’ve done some investigations :)
> >
> > First of all, as soon as we see cache-dependency, the problem should not be with .js-code, it’s recompiled, but as far as I can see Angular uses caching.
> >
> > The problem looks like a result of a change in file guacTiledClients.html. There was a line:
> > ng-style="{ 'width' : getTileWidth(), 'height' : getTileHeight() }”
> >
> > And it was working perfectly, but in a new version we have:
> > ng-attr-style="{{ getTileGridCSS(clientGroup, $index) }}”
> >
> > So, when viewing Elements in Safari, I can see the following structure:
> >
> > <guac-tiled-clients ...>
> > <ul ...>
> > <!-- ngRepeat: client in clientGroup.clients —>
> > <li class="client-tile ng-scope focused" ...>
> > </li>
> > <!-- end ngRepeat: client in clientGroup.clients —>
> > </ul>
> > </guac-tiled-clients>
> >
> > And magically inner li element has non-zero width, but zero height. If I change it manually, the picture appears, and all is working good.
> >
> > Digging deeper shows me the source of zero height - it is mentioned in client.css, with two classes:
> > .client-view
> > .client-view-content
> >
> > Both of them declare “100%” height, but looks like it does not work in Safari. Any ideas?
> >
> > Cheers,
> > Leo
> >
> >> On 5 Aug 2021, at 23:49, Mike Jumper <mi...@glyptodon.com> wrote:
> >>
> >> Well, so far no luck reproducing what you're seeing, but perhaps my test setup is too up-to-date:
> >>
> >> macOS 11.5
> >> Safari 14.1.2
> >>
> >> I'll try again and, failing that, see if there's some way to duplicate your exact setup down to the version.
> >>
> >> Michael Jumper
> >> CEO, Lead Developer
> >> Glyptodon Inc.
> >>
> >>
> >> On Wed, Aug 4, 2021 at 1:49 PM Mike Jumper <mi...@glyptodon.com> wrote:
> >> Ah, OK.
> >>
> >> On Wed, Aug 4, 2021, 13:48 Leo Nikolaev <ni...@kasatkina.org> wrote:
> >> I do need cherry-pick to fix regression from 641, you’ve told me that; I’ve found it fixed on 037ed212 and put it on. Without it Guacamole fails to even show me a login form, complaining about MySQL syntax errors.
> >>
> >> Cheers,
> >> Leo
> >>
> >>> On 4 Aug 2021, at 23:45, Mike Jumper <mi...@glyptodon.com> wrote:
> >>>
> >>> On Wed, Aug 4, 2021, 13:33 Leo Nikolaev <ni...@kasatkina.org> wrote:
> >>> Okay, I found it.
> >>>
> >>> Safari bug is in the GUACAMOLE-724 series of commits.
> >>> Broken commit: c2b2522 - GUACAMOLE-724: Correct rendering of tiled clients on IE10+ by migrating to grid layout.
> >>>
> >>> Unfortunately, I believe the code from that commit has already been replaced with a different approach.
> >>>
> >>> Last working commit: 09288b7
> >>>
> >>> This may be helpful.
> >>>
> >>>
> >>> Steps to reproduce:
> >>>
> >>> 1. Checkout commit:
> >>>
> >>> git checkout c2b2522
> >>>
> >>> 2. Cherry-pick fix from GUACAMOLE-641:
> >>>
> >>> git cherry-pick -m 1 037ed212
> >>>
> >>> What does cherry-picking this commit have to do with the issue at hand? Are you unable to reproduce the issue unless you cherry pick this?
> >>>
> >>> ...
> >>> Should I make a video?
> >>>
> >>> No, I don't think that would help. We need to be able reproduce what you're seeing ourselves.
> >>>
> >>> The steps to reproduce (connect to something) are straightforward; it just so far is working absolutely fine when I test against Safari myself.
> >>>
> >>> - Mike
> >>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> >> For additional commands, e-mail: user-help@guacamole.apache.org
> >>
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> > For additional commands, e-mail: user-help@guacamole.apache.org
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> For additional commands, e-mail: user-help@guacamole.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
For additional commands, e-mail: user-help@guacamole.apache.org
Re: guacamole broken on Safari
Posted by Mike Jumper <mi...@glyptodon.com>.
Nope, not as of yet. Based on what you're seeing, it sounds like a CSS
issue, yes, but not an issue with grid layout. The code that you're running
across as the cause was only an intermediate attempt during development and
actually isn't present any longer. The layout was migrated to flexbox prior
to merge.
That said, I would expect something CSS-related to be consistent and
abundantly reproducible. So far, I can't reproduce it at all in the only
copy of Safari I (currently) have access to.
On Wed, Aug 11, 2021 at 2:31 PM Leo Nikolaev <ni...@kasatkina.org> wrote:
> Mike, any update on this? Looks like a CSS bug in tile grid :)
>
> Cheers,
> Leo
>
> > On 6 Aug 2021, at 10:42, Leo Nikolaev <ni...@kasatkina.org> wrote:
> >
> > Okay, I’ve done some investigations :)
> >
> > First of all, as soon as we see cache-dependency, the problem should not
> be with .js-code, it’s recompiled, but as far as I can see Angular uses
> caching.
> >
> > The problem looks like a result of a change in file
> guacTiledClients.html. There was a line:
> > ng-style="{ 'width' : getTileWidth(), 'height' : getTileHeight() }”
> >
> > And it was working perfectly, but in a new version we have:
> > ng-attr-style="{{ getTileGridCSS(clientGroup, $index) }}”
> >
> > So, when viewing Elements in Safari, I can see the following structure:
> >
> > <guac-tiled-clients ...>
> > <ul ...>
> > <!-- ngRepeat: client in clientGroup.clients —>
> > <li class="client-tile ng-scope focused" ...>
> > </li>
> > <!-- end ngRepeat: client in clientGroup.clients —>
> > </ul>
> > </guac-tiled-clients>
> >
> > And magically inner li element has non-zero width, but zero height. If I
> change it manually, the picture appears, and all is working good.
> >
> > Digging deeper shows me the source of zero height - it is mentioned in
> client.css, with two classes:
> > .client-view
> > .client-view-content
> >
> > Both of them declare “100%” height, but looks like it does not work in
> Safari. Any ideas?
> >
> > Cheers,
> > Leo
> >
> >> On 5 Aug 2021, at 23:49, Mike Jumper <mi...@glyptodon.com> wrote:
> >>
> >> Well, so far no luck reproducing what you're seeing, but perhaps my
> test setup is too up-to-date:
> >>
> >> macOS 11.5
> >> Safari 14.1.2
> >>
> >> I'll try again and, failing that, see if there's some way to duplicate
> your exact setup down to the version.
> >>
> >> Michael Jumper
> >> CEO, Lead Developer
> >> Glyptodon Inc.
> >>
> >>
> >> On Wed, Aug 4, 2021 at 1:49 PM Mike Jumper <mi...@glyptodon.com>
> wrote:
> >> Ah, OK.
> >>
> >> On Wed, Aug 4, 2021, 13:48 Leo Nikolaev <ni...@kasatkina.org> wrote:
> >> I do need cherry-pick to fix regression from 641, you’ve told me that;
> I’ve found it fixed on 037ed212 and put it on. Without it Guacamole fails
> to even show me a login form, complaining about MySQL syntax errors.
> >>
> >> Cheers,
> >> Leo
> >>
> >>> On 4 Aug 2021, at 23:45, Mike Jumper <mi...@glyptodon.com>
> wrote:
> >>>
> >>> On Wed, Aug 4, 2021, 13:33 Leo Nikolaev <ni...@kasatkina.org>
> wrote:
> >>> Okay, I found it.
> >>>
> >>> Safari bug is in the GUACAMOLE-724 series of commits.
> >>> Broken commit: c2b2522 - GUACAMOLE-724: Correct rendering of tiled
> clients on IE10+ by migrating to grid layout.
> >>>
> >>> Unfortunately, I believe the code from that commit has already been
> replaced with a different approach.
> >>>
> >>> Last working commit: 09288b7
> >>>
> >>> This may be helpful.
> >>>
> >>>
> >>> Steps to reproduce:
> >>>
> >>> 1. Checkout commit:
> >>>
> >>> git checkout c2b2522
> >>>
> >>> 2. Cherry-pick fix from GUACAMOLE-641:
> >>>
> >>> git cherry-pick -m 1 037ed212
> >>>
> >>> What does cherry-picking this commit have to do with the issue at
> hand? Are you unable to reproduce the issue unless you cherry pick this?
> >>>
> >>> ...
> >>> Should I make a video?
> >>>
> >>> No, I don't think that would help. We need to be able reproduce what
> you're seeing ourselves.
> >>>
> >>> The steps to reproduce (connect to something) are straightforward; it
> just so far is working absolutely fine when I test against Safari myself.
> >>>
> >>> - Mike
> >>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> >> For additional commands, e-mail: user-help@guacamole.apache.org
> >>
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> > For additional commands, e-mail: user-help@guacamole.apache.org
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> For additional commands, e-mail: user-help@guacamole.apache.org
>
>
Re: guacamole broken on Safari
Posted by Leo Nikolaev <ni...@kasatkina.org>.
Mike, any update on this? Looks like a CSS bug in tile grid :)
Cheers,
Leo
> On 6 Aug 2021, at 10:42, Leo Nikolaev <ni...@kasatkina.org> wrote:
>
> Okay, I’ve done some investigations :)
>
> First of all, as soon as we see cache-dependency, the problem should not be with .js-code, it’s recompiled, but as far as I can see Angular uses caching.
>
> The problem looks like a result of a change in file guacTiledClients.html. There was a line:
> ng-style="{ 'width' : getTileWidth(), 'height' : getTileHeight() }”
>
> And it was working perfectly, but in a new version we have:
> ng-attr-style="{{ getTileGridCSS(clientGroup, $index) }}”
>
> So, when viewing Elements in Safari, I can see the following structure:
>
> <guac-tiled-clients ...>
> <ul ...>
> <!-- ngRepeat: client in clientGroup.clients —>
> <li class="client-tile ng-scope focused" ...>
> </li>
> <!-- end ngRepeat: client in clientGroup.clients —>
> </ul>
> </guac-tiled-clients>
>
> And magically inner li element has non-zero width, but zero height. If I change it manually, the picture appears, and all is working good.
>
> Digging deeper shows me the source of zero height - it is mentioned in client.css, with two classes:
> .client-view
> .client-view-content
>
> Both of them declare “100%” height, but looks like it does not work in Safari. Any ideas?
>
> Cheers,
> Leo
>
>> On 5 Aug 2021, at 23:49, Mike Jumper <mi...@glyptodon.com> wrote:
>>
>> Well, so far no luck reproducing what you're seeing, but perhaps my test setup is too up-to-date:
>>
>> macOS 11.5
>> Safari 14.1.2
>>
>> I'll try again and, failing that, see if there's some way to duplicate your exact setup down to the version.
>>
>> Michael Jumper
>> CEO, Lead Developer
>> Glyptodon Inc.
>>
>>
>> On Wed, Aug 4, 2021 at 1:49 PM Mike Jumper <mi...@glyptodon.com> wrote:
>> Ah, OK.
>>
>> On Wed, Aug 4, 2021, 13:48 Leo Nikolaev <ni...@kasatkina.org> wrote:
>> I do need cherry-pick to fix regression from 641, you’ve told me that; I’ve found it fixed on 037ed212 and put it on. Without it Guacamole fails to even show me a login form, complaining about MySQL syntax errors.
>>
>> Cheers,
>> Leo
>>
>>> On 4 Aug 2021, at 23:45, Mike Jumper <mi...@glyptodon.com> wrote:
>>>
>>> On Wed, Aug 4, 2021, 13:33 Leo Nikolaev <ni...@kasatkina.org> wrote:
>>> Okay, I found it.
>>>
>>> Safari bug is in the GUACAMOLE-724 series of commits.
>>> Broken commit: c2b2522 - GUACAMOLE-724: Correct rendering of tiled clients on IE10+ by migrating to grid layout.
>>>
>>> Unfortunately, I believe the code from that commit has already been replaced with a different approach.
>>>
>>> Last working commit: 09288b7
>>>
>>> This may be helpful.
>>>
>>>
>>> Steps to reproduce:
>>>
>>> 1. Checkout commit:
>>>
>>> git checkout c2b2522
>>>
>>> 2. Cherry-pick fix from GUACAMOLE-641:
>>>
>>> git cherry-pick -m 1 037ed212
>>>
>>> What does cherry-picking this commit have to do with the issue at hand? Are you unable to reproduce the issue unless you cherry pick this?
>>>
>>> ...
>>> Should I make a video?
>>>
>>> No, I don't think that would help. We need to be able reproduce what you're seeing ourselves.
>>>
>>> The steps to reproduce (connect to something) are straightforward; it just so far is working absolutely fine when I test against Safari myself.
>>>
>>> - Mike
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
>> For additional commands, e-mail: user-help@guacamole.apache.org
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> For additional commands, e-mail: user-help@guacamole.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
For additional commands, e-mail: user-help@guacamole.apache.org
Re: guacamole broken on Safari
Posted by Leo Nikolaev <ni...@kasatkina.org>.
Okay, I’ve done some investigations :)
First of all, as soon as we see cache-dependency, the problem should not be with .js-code, it’s recompiled, but as far as I can see Angular uses caching.
The problem looks like a result of a change in file guacTiledClients.html. There was a line:
ng-style="{ 'width' : getTileWidth(), 'height' : getTileHeight() }”
And it was working perfectly, but in a new version we have:
ng-attr-style="{{ getTileGridCSS(clientGroup, $index) }}”
So, when viewing Elements in Safari, I can see the following structure:
<guac-tiled-clients ...>
<ul ...>
<!-- ngRepeat: client in clientGroup.clients —>
<li class="client-tile ng-scope focused" ...>
</li>
<!-- end ngRepeat: client in clientGroup.clients —>
</ul>
</guac-tiled-clients>
And magically inner li element has non-zero width, but zero height. If I change it manually, the picture appears, and all is working good.
Digging deeper shows me the source of zero height - it is mentioned in client.css, with two classes:
.client-view
.client-view-content
Both of them declare “100%” height, but looks like it does not work in Safari. Any ideas?
Cheers,
Leo
> On 5 Aug 2021, at 23:49, Mike Jumper <mi...@glyptodon.com> wrote:
>
> Well, so far no luck reproducing what you're seeing, but perhaps my test setup is too up-to-date:
>
> macOS 11.5
> Safari 14.1.2
>
> I'll try again and, failing that, see if there's some way to duplicate your exact setup down to the version.
>
> Michael Jumper
> CEO, Lead Developer
> Glyptodon Inc.
>
>
> On Wed, Aug 4, 2021 at 1:49 PM Mike Jumper <mi...@glyptodon.com> wrote:
> Ah, OK.
>
> On Wed, Aug 4, 2021, 13:48 Leo Nikolaev <ni...@kasatkina.org> wrote:
> I do need cherry-pick to fix regression from 641, you’ve told me that; I’ve found it fixed on 037ed212 and put it on. Without it Guacamole fails to even show me a login form, complaining about MySQL syntax errors.
>
> Cheers,
> Leo
>
> > On 4 Aug 2021, at 23:45, Mike Jumper <mi...@glyptodon.com> wrote:
> >
> > On Wed, Aug 4, 2021, 13:33 Leo Nikolaev <ni...@kasatkina.org> wrote:
> > Okay, I found it.
> >
> > Safari bug is in the GUACAMOLE-724 series of commits.
> > Broken commit: c2b2522 - GUACAMOLE-724: Correct rendering of tiled clients on IE10+ by migrating to grid layout.
> >
> > Unfortunately, I believe the code from that commit has already been replaced with a different approach.
> >
> > Last working commit: 09288b7
> >
> > This may be helpful.
> >
> >
> > Steps to reproduce:
> >
> > 1. Checkout commit:
> >
> > git checkout c2b2522
> >
> > 2. Cherry-pick fix from GUACAMOLE-641:
> >
> > git cherry-pick -m 1 037ed212
> >
> > What does cherry-picking this commit have to do with the issue at hand? Are you unable to reproduce the issue unless you cherry pick this?
> >
> > ...
> > Should I make a video?
> >
> > No, I don't think that would help. We need to be able reproduce what you're seeing ourselves.
> >
> > The steps to reproduce (connect to something) are straightforward; it just so far is working absolutely fine when I test against Safari myself.
> >
> > - Mike
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> For additional commands, e-mail: user-help@guacamole.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
For additional commands, e-mail: user-help@guacamole.apache.org
Re: guacamole broken on Safari
Posted by Mike Jumper <mi...@glyptodon.com>.
Well, so far no luck reproducing what you're seeing, but perhaps my test
setup is too up-to-date:
macOS 11.5
Safari 14.1.2
I'll try again and, failing that, see if there's some way to duplicate your
exact setup down to the version.
Michael Jumper
CEO, Lead Developer
Glyptodon Inc <https://glyp.to/>.
On Wed, Aug 4, 2021 at 1:49 PM Mike Jumper <mi...@glyptodon.com>
wrote:
> Ah, OK.
>
> On Wed, Aug 4, 2021, 13:48 Leo Nikolaev <ni...@kasatkina.org> wrote:
>
>> I do need cherry-pick to fix regression from 641, you’ve told me that;
>> I’ve found it fixed on 037ed212 and put it on. Without it Guacamole fails
>> to even show me a login form, complaining about MySQL syntax errors.
>>
>> Cheers,
>> Leo
>>
>> > On 4 Aug 2021, at 23:45, Mike Jumper <mi...@glyptodon.com> wrote:
>> >
>> > On Wed, Aug 4, 2021, 13:33 Leo Nikolaev <ni...@kasatkina.org> wrote:
>> > Okay, I found it.
>> >
>> > Safari bug is in the GUACAMOLE-724 series of commits.
>> > Broken commit: c2b2522 - GUACAMOLE-724: Correct rendering of tiled
>> clients on IE10+ by migrating to grid layout.
>> >
>> > Unfortunately, I believe the code from that commit has already been
>> replaced with a different approach.
>> >
>> > Last working commit: 09288b7
>> >
>> > This may be helpful.
>> >
>> >
>> > Steps to reproduce:
>> >
>> > 1. Checkout commit:
>> >
>> > git checkout c2b2522
>> >
>> > 2. Cherry-pick fix from GUACAMOLE-641:
>> >
>> > git cherry-pick -m 1 037ed212
>> >
>> > What does cherry-picking this commit have to do with the issue at hand?
>> Are you unable to reproduce the issue unless you cherry pick this?
>> >
>> > ...
>> > Should I make a video?
>> >
>> > No, I don't think that would help. We need to be able reproduce what
>> you're seeing ourselves.
>> >
>> > The steps to reproduce (connect to something) are straightforward; it
>> just so far is working absolutely fine when I test against Safari myself.
>> >
>> > - Mike
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
>> For additional commands, e-mail: user-help@guacamole.apache.org
>>
>>
Re: guacamole broken on Safari
Posted by Mike Jumper <mi...@glyptodon.com>.
Ah, OK.
On Wed, Aug 4, 2021, 13:48 Leo Nikolaev <ni...@kasatkina.org> wrote:
> I do need cherry-pick to fix regression from 641, you’ve told me that;
> I’ve found it fixed on 037ed212 and put it on. Without it Guacamole fails
> to even show me a login form, complaining about MySQL syntax errors.
>
> Cheers,
> Leo
>
> > On 4 Aug 2021, at 23:45, Mike Jumper <mi...@glyptodon.com> wrote:
> >
> > On Wed, Aug 4, 2021, 13:33 Leo Nikolaev <ni...@kasatkina.org> wrote:
> > Okay, I found it.
> >
> > Safari bug is in the GUACAMOLE-724 series of commits.
> > Broken commit: c2b2522 - GUACAMOLE-724: Correct rendering of tiled
> clients on IE10+ by migrating to grid layout.
> >
> > Unfortunately, I believe the code from that commit has already been
> replaced with a different approach.
> >
> > Last working commit: 09288b7
> >
> > This may be helpful.
> >
> >
> > Steps to reproduce:
> >
> > 1. Checkout commit:
> >
> > git checkout c2b2522
> >
> > 2. Cherry-pick fix from GUACAMOLE-641:
> >
> > git cherry-pick -m 1 037ed212
> >
> > What does cherry-picking this commit have to do with the issue at hand?
> Are you unable to reproduce the issue unless you cherry pick this?
> >
> > ...
> > Should I make a video?
> >
> > No, I don't think that would help. We need to be able reproduce what
> you're seeing ourselves.
> >
> > The steps to reproduce (connect to something) are straightforward; it
> just so far is working absolutely fine when I test against Safari myself.
> >
> > - Mike
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> For additional commands, e-mail: user-help@guacamole.apache.org
>
>
Re: guacamole broken on Safari
Posted by Leo Nikolaev <ni...@kasatkina.org>.
I do need cherry-pick to fix regression from 641, you’ve told me that; I’ve found it fixed on 037ed212 and put it on. Without it Guacamole fails to even show me a login form, complaining about MySQL syntax errors.
Cheers,
Leo
> On 4 Aug 2021, at 23:45, Mike Jumper <mi...@glyptodon.com> wrote:
>
> On Wed, Aug 4, 2021, 13:33 Leo Nikolaev <ni...@kasatkina.org> wrote:
> Okay, I found it.
>
> Safari bug is in the GUACAMOLE-724 series of commits.
> Broken commit: c2b2522 - GUACAMOLE-724: Correct rendering of tiled clients on IE10+ by migrating to grid layout.
>
> Unfortunately, I believe the code from that commit has already been replaced with a different approach.
>
> Last working commit: 09288b7
>
> This may be helpful.
>
>
> Steps to reproduce:
>
> 1. Checkout commit:
>
> git checkout c2b2522
>
> 2. Cherry-pick fix from GUACAMOLE-641:
>
> git cherry-pick -m 1 037ed212
>
> What does cherry-picking this commit have to do with the issue at hand? Are you unable to reproduce the issue unless you cherry pick this?
>
> ...
> Should I make a video?
>
> No, I don't think that would help. We need to be able reproduce what you're seeing ourselves.
>
> The steps to reproduce (connect to something) are straightforward; it just so far is working absolutely fine when I test against Safari myself.
>
> - Mike
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
For additional commands, e-mail: user-help@guacamole.apache.org
Re: guacamole broken on Safari
Posted by Mike Jumper <mi...@glyptodon.com>.
On Wed, Aug 4, 2021, 13:33 Leo Nikolaev <ni...@kasatkina.org> wrote:
> Okay, I found it.
>
> Safari bug is in the GUACAMOLE-724 series of commits.
> Broken commit: c2b2522 - GUACAMOLE-724: Correct rendering of tiled clients
> on IE10+ by migrating to grid layout.
>
Unfortunately, I believe the code from that commit has already been
replaced with a different approach.
Last working commit: 09288b7
>
This may be helpful.
> Steps to reproduce:
>
> 1. Checkout commit:
>
> git checkout c2b2522
>
> 2. Cherry-pick fix from GUACAMOLE-641:
>
> git cherry-pick -m 1 037ed212
>
What does cherry-picking this commit have to do with the issue at hand? Are
you unable to reproduce the issue unless you cherry pick this?
...
> Should I make a video?
>
No, I don't think that would help. We need to be able reproduce what you're
seeing ourselves.
The steps to reproduce (connect to something) are straightforward; it just
so far is working absolutely fine when I test against Safari myself.
- Mike
Re: guacamole broken on Safari
Posted by Leo Nikolaev <ni...@kasatkina.org>.
Okay, I found it.
Safari bug is in the GUACAMOLE-724 series of commits.
Broken commit: c2b2522 - GUACAMOLE-724: Correct rendering of tiled clients on IE10+ by migrating to grid layout.
Last working commit: 09288b7
Steps to reproduce:
1. Checkout commit:
git checkout c2b2522
2. Cherry-pick fix from GUACAMOLE-641:
git cherry-pick -m 1 037ed212
3. Build docker image:
docker build -t guacamole/guacamole:c2b2522 .
4. Run it any way you like.
5. IMPORTANT: Clear Safari cache.
Should I make a video?
Cheers,
Leo
> On 4 Aug 2021, at 13:25, Mike Jumper <mi...@glyptodon.com> wrote:
>
> On Wed, Aug 4, 2021 at 1:36 AM Leo Nikolaev <ni...@kasatkina.org> wrote:
> Ha, I found an interesting issue about the bug.
>
> If you run a working guacamole, it puts something in your Safari’s browser cache to keep it working, even if future builds have a bug. Cleaning up the cache fires up a bug. I’m still bisecting it, but I’m close.
>
> It's not so much that Guacamole puts something in the browser cache, but that the browser caches Guacamole's JavaScript, CSS, etc. If the issue you're seeing is due to the contents of a resource cached by the browser, then that behavior will end up being conditional on whether the issue happens to still be in the cache.
>
> Depending on how frequently the contents of the cache affected the results of the test, the final result of the bisect may end up being inconsistent, with git believing that some good commits are bad, some bad commits are good, etc.
>
> - Mike
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
For additional commands, e-mail: user-help@guacamole.apache.org
Re: guacamole broken on Safari
Posted by Mike Jumper <mi...@glyptodon.com>.
On Wed, Aug 4, 2021 at 1:36 AM Leo Nikolaev <ni...@kasatkina.org> wrote:
> Ha, I found an interesting issue about the bug.
>
> If you run a working guacamole, it puts something in your Safari’s browser
> cache to keep it working, even if future builds have a bug. Cleaning up the
> cache fires up a bug. I’m still bisecting it, but I’m close.
It's not so much that Guacamole *puts* something in the browser cache, but
that the browser caches Guacamole's JavaScript, CSS, etc. If the issue
you're seeing is due to the contents of a resource cached by the browser,
then that behavior will end up being conditional on whether the issue
happens to still be in the cache.
Depending on how frequently the contents of the cache affected the results
of the test, the final result of the bisect may end up being
inconsistent, with git believing that some good commits are bad, some bad
commits are good, etc.
- Mike
Re: guacamole broken on Safari
Posted by Leo Nikolaev <ni...@kasatkina.org>.
Ha, I found an interesting issue about the bug.
If you run a working guacamole, it puts something in your Safari’s browser cache to keep it working, even if future builds have a bug. Cleaning up the cache fires up a bug. I’m still bisecting it, but I’m close.
Cheers,
Leo
> On 3 Aug 2021, at 23:52, Mike Jumper <mi...@glyptodon.com> wrote:
>
> Yes, it was a regression introduced during development of GUACAMOLE-641 and was subsequently fixed.
>
> If you run into an issue that prevents testing entirely (can't build, can't login, etc.), the best way to move beyond that would be to skip that commit with "git bisect skip" so that git knows that the status of that commit can't be determined. It'll continue its binary search and avoid the vicinity of that commit unless necessary.
>
> Michael Jumper
> CEO, Lead Developer
> Glyptodon Inc.
>
>
> On Tue, Aug 3, 2021 at 1:27 PM Leo Nikolaev <ni...@kasatkina.org> wrote:
> So, now I have a very nasty bug: some commits do fail to even show the login page of Guacamole with some weird MySQL error:
>
> ### The error may exist in org/apache/guacamole/auth/jdbc/user/UserMapper.xml
> ### The error may involve defaultParameterMap
> ### The error occurred while setting parameters
> ### SQL: SELECT guacamole_user.user_id, guacamole_entity.entity_id, guacamole_entity.name, password_hash, password_salt, password_date, disabled, expired, access_window_start, access_window_end, valid_from, valid_until, timezone, full_name, email_address, organization, organizational_role, MAX(start_date) AS last_active FROM guacamole_user JOIN guacamole_entity ON guacamole_user.entity_id = guacamole_entity.entity_id LEFT JOIN guacamole_user_history ON guacamole_user_history.user_id = guacamole_user.user_id WHERE guacamole_entity.name = ? AND guacamole_entity.type = 'USER' GROUP BY guacamole_user.user_id, guacamole_entity.entity_id; SELECT guacamole_user_attribute.user_id, guacamole_user_attribute.attribute_name, guacamole_user_attribute.attribute_value FROM guacamole_user_attribute JOIN guacamole_user ON guacamole_user.user_id = guacamole_user_attribute.user_id JOIN guacamole_entity ON guacamole_user.entity_id = guacamole_entity.entity_id WHERE guacamole_entity.name = ? AND guacamole_entity.type = 'USER'
> ### Cause: com.mysql.jdbc.exceptions.jdbc4.MySQLSyntaxErrorException: You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'SELECT
> guacamole_user_attribute.user_id,
> guacamole_user_' at line 28
>
> Have you seen anything like this? I don’t see any changes in MySQL schemas, but commits 448ebb5 and 2b8eb44 have this bug. Commit 037ed21 and c239b6e do not have this bug at all.
>
> Cheers,
> Leo
>
> > On 3 Aug 2021, at 00:11, Mike Jumper <mi...@glyptodon.com> wrote:
> >
> > On Mon, Aug 2, 2021, 14:07 Leo Nikolaev <ni...@kasatkina.org> wrote:
> > Erm, I can’t do bisect, I just build it and check manually :)
> >
> > Bisect will save you a *ton* of time while reducing the number of checks. You'll be able to cover the full relevant span of commits in O(log n).
> >
> > Btw, I am using OIDC auth via Keycloak, but I’m sure it's not the reason.
> >
> > Probably true, but the fewer variables and assumptions, the better.
> >
> > - Mike
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> For additional commands, e-mail: user-help@guacamole.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
For additional commands, e-mail: user-help@guacamole.apache.org
Re: guacamole broken on Safari
Posted by Mike Jumper <mi...@glyptodon.com>.
Yes, it was a regression introduced during development of GUACAMOLE-641 and
was subsequently fixed.
If you run into an issue that prevents testing entirely (can't build, can't
login, etc.), the best way to move beyond that would be to skip that commit
with "git bisect skip" so that git knows that the status of that commit
can't be determined. It'll continue its binary search and avoid the
vicinity of that commit unless necessary.
Michael Jumper
CEO, Lead Developer
Glyptodon Inc <https://glyp.to/>.
On Tue, Aug 3, 2021 at 1:27 PM Leo Nikolaev <ni...@kasatkina.org> wrote:
> So, now I have a very nasty bug: some commits do fail to even show the
> login page of Guacamole with some weird MySQL error:
>
> ### The error may exist in
> org/apache/guacamole/auth/jdbc/user/UserMapper.xml
> ### The error may involve defaultParameterMap
> ### The error occurred while setting parameters
> ### SQL: SELECT guacamole_user.user_id,
> guacamole_entity.entity_id, guacamole_entity.name,
> password_hash, password_salt, password_date,
> disabled, expired, access_window_start,
> access_window_end, valid_from,
> valid_until, timezone, full_name,
> email_address, organization, organizational_role,
> MAX(start_date) AS last_active FROM guacamole_user
> JOIN guacamole_entity ON guacamole_user.entity_id =
> guacamole_entity.entity_id LEFT JOIN guacamole_user_history ON
> guacamole_user_history.user_id = guacamole_user.user_id WHERE
> guacamole_entity.name = ? AND guacamole_entity.type =
> 'USER' GROUP BY guacamole_user.user_id,
> guacamole_entity.entity_id; SELECT
> guacamole_user_attribute.user_id,
> guacamole_user_attribute.attribute_name,
> guacamole_user_attribute.attribute_value FROM
> guacamole_user_attribute JOIN guacamole_user ON
> guacamole_user.user_id = guacamole_user_attribute.user_id JOIN
> guacamole_entity ON guacamole_user.entity_id = guacamole_entity.entity_id
> WHERE guacamole_entity.name = ? AND
> guacamole_entity.type = 'USER'
> ### Cause: com.mysql.jdbc.exceptions.jdbc4.MySQLSyntaxErrorException: You
> have an error in your SQL syntax; check the manual that corresponds to your
> MySQL server version for the right syntax to use near 'SELECT
> guacamole_user_attribute.user_id,
> guacamole_user_' at line 28
>
> Have you seen anything like this? I don’t see any changes in MySQL
> schemas, but commits 448ebb5 and 2b8eb44 have this bug. Commit 037ed21 and
> c239b6e do not have this bug at all.
>
> Cheers,
> Leo
>
> > On 3 Aug 2021, at 00:11, Mike Jumper <mi...@glyptodon.com> wrote:
> >
> > On Mon, Aug 2, 2021, 14:07 Leo Nikolaev <ni...@kasatkina.org> wrote:
> > Erm, I can’t do bisect, I just build it and check manually :)
> >
> > Bisect will save you a *ton* of time while reducing the number of
> checks. You'll be able to cover the full relevant span of commits in O(log
> n).
> >
> > Btw, I am using OIDC auth via Keycloak, but I’m sure it's not the reason.
> >
> > Probably true, but the fewer variables and assumptions, the better.
> >
> > - Mike
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> For additional commands, e-mail: user-help@guacamole.apache.org
>
>
Re: guacamole broken on Safari
Posted by Leo Nikolaev <ni...@kasatkina.org>.
So, now I have a very nasty bug: some commits do fail to even show the login page of Guacamole with some weird MySQL error:
### The error may exist in org/apache/guacamole/auth/jdbc/user/UserMapper.xml
### The error may involve defaultParameterMap
### The error occurred while setting parameters
### SQL: SELECT guacamole_user.user_id, guacamole_entity.entity_id, guacamole_entity.name, password_hash, password_salt, password_date, disabled, expired, access_window_start, access_window_end, valid_from, valid_until, timezone, full_name, email_address, organization, organizational_role, MAX(start_date) AS last_active FROM guacamole_user JOIN guacamole_entity ON guacamole_user.entity_id = guacamole_entity.entity_id LEFT JOIN guacamole_user_history ON guacamole_user_history.user_id = guacamole_user.user_id WHERE guacamole_entity.name = ? AND guacamole_entity.type = 'USER' GROUP BY guacamole_user.user_id, guacamole_entity.entity_id; SELECT guacamole_user_attribute.user_id, guacamole_user_attribute.attribute_name, guacamole_user_attribute.attribute_value FROM guacamole_user_attribute JOIN guacamole_user ON guacamole_user.user_id = guacamole_user_attribute.user_id JOIN guacamole_entity ON guacamole_user.entity_id = guacamole_entity.entity_id WHERE guacamole_entity.name = ? AND guacamole_entity.type = 'USER'
### Cause: com.mysql.jdbc.exceptions.jdbc4.MySQLSyntaxErrorException: You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'SELECT
guacamole_user_attribute.user_id,
guacamole_user_' at line 28
Have you seen anything like this? I don’t see any changes in MySQL schemas, but commits 448ebb5 and 2b8eb44 have this bug. Commit 037ed21 and c239b6e do not have this bug at all.
Cheers,
Leo
> On 3 Aug 2021, at 00:11, Mike Jumper <mi...@glyptodon.com> wrote:
>
> On Mon, Aug 2, 2021, 14:07 Leo Nikolaev <ni...@kasatkina.org> wrote:
> Erm, I can’t do bisect, I just build it and check manually :)
>
> Bisect will save you a *ton* of time while reducing the number of checks. You'll be able to cover the full relevant span of commits in O(log n).
>
> Btw, I am using OIDC auth via Keycloak, but I’m sure it's not the reason.
>
> Probably true, but the fewer variables and assumptions, the better.
>
> - Mike
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
For additional commands, e-mail: user-help@guacamole.apache.org
Re: guacamole broken on Safari
Posted by Mike Jumper <mi...@glyptodon.com>.
On Mon, Aug 2, 2021, 14:07 Leo Nikolaev <ni...@kasatkina.org> wrote:
> Erm, I can’t do bisect, I just build it and check manually :)
>
Bisect will save you a *ton* of time while reducing the number of checks.
You'll be able to cover the full relevant span of commits in O(log n).
Btw, I am using OIDC auth via Keycloak, but I’m sure it's not the reason.
Probably true, but the fewer variables and assumptions, the better.
- Mike
Re: guacamole broken on Safari
Posted by Leo Nikolaev <ni...@kasatkina.org>.
Erm, I can’t do bisect, I just build it and check manually :)
Btw, I am using OIDC auth via Keycloak, but I’m sure it's not the reason.
> On 3 Aug 2021, at 00:04, Mike Jumper <mi...@glyptodon.com> wrote:
>
> Is this from a git bisect?
>
> - Mike
>
> On Mon, Aug 2, 2021, 13:59 Leo Nikolaev <ni...@kasatkina.org> wrote:
> Okay, I’ve done some triaging by building and checking manually some commits.
>
> Commit 037ed21 is definetely okay, and it even has the logout support, so fits perfectly.
>
> There is a bunch of commits at 8th of July then, and the last of them, dd86130 is completely broken.
>
> Commit cc239b6e has the bug with Safari.
>
> So I think bug was introduced somewhere at 8th of July, will try to test more commits.
>
> Cheers,
> Leo
>
> > On 30 Jul 2021, at 04:00, Mike Jumper <mi...@glyptodon.com> wrote:
> >
> > On Wed, Jul 28, 2021 at 3:36 AM Leo Nikolaev <ni...@kasatkina.org> wrote:
> > If comfortable, I can record a screencast or do a Zoom demo of what am I doing.
> >
> > I’m absolutely sure about latest master - I mount it inside Docker-container, and when I remove -v parameter it all works perfectly on 1.3.0.
> >
> > docker run --name guacamole \
> > --link guacd \
> > --link mysql \
> > -v $(pwd)/guacamole:/etc/guacamole \
> > -v $(pwd)/guacamole-client/guacamole/target/guacamole-1.3.0.war:/opt/guacamole/guacamole.war \
> > -e MYSQL_DATABASE=guacamole \
> > -e MYSQL_USER=root \
> > -e MYSQL_PASSWORD=guacamole \
> > -e GUACAMOLE_HOME=/etc/guacamole \
> > -d -p 8080:8080 \
> > guacamole/guacamole
> >
> > Building it from repo, yes.
> >
> > Manually layering an externally-built, newer .war file over the .war within the Docker image from a previous release is a bit funky. To ensure things are sane, can you build the Docker image directly from the same source and get rid of the volume mount? The Dockerfile for the image is part of the guacamole-client repository, so you can just do a "docker build" from the root of the source that you cloned from git to produce an updated image.
> >
> > Thanks,
> >
> > Michael Jumper
> > CEO, Lead Developer
> > Glyptodon Inc.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> For additional commands, e-mail: user-help@guacamole.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
For additional commands, e-mail: user-help@guacamole.apache.org
Re: guacamole broken on Safari
Posted by Mike Jumper <mi...@glyptodon.com>.
Is this from a git bisect?
- Mike
On Mon, Aug 2, 2021, 13:59 Leo Nikolaev <ni...@kasatkina.org> wrote:
> Okay, I’ve done some triaging by building and checking manually some
> commits.
>
> Commit 037ed21 is definetely okay, and it even has the logout support, so
> fits perfectly.
>
> There is a bunch of commits at 8th of July then, and the last of them,
> dd86130 is completely broken.
>
> Commit cc239b6e has the bug with Safari.
>
> So I think bug was introduced somewhere at 8th of July, will try to test
> more commits.
>
> Cheers,
> Leo
>
> > On 30 Jul 2021, at 04:00, Mike Jumper <mi...@glyptodon.com> wrote:
> >
> > On Wed, Jul 28, 2021 at 3:36 AM Leo Nikolaev <ni...@kasatkina.org>
> wrote:
> > If comfortable, I can record a screencast or do a Zoom demo of what am I
> doing.
> >
> > I’m absolutely sure about latest master - I mount it inside
> Docker-container, and when I remove -v parameter it all works perfectly on
> 1.3.0.
> >
> > docker run --name guacamole \
> > --link guacd \
> > --link mysql \
> > -v $(pwd)/guacamole:/etc/guacamole \
> > -v
> $(pwd)/guacamole-client/guacamole/target/guacamole-1.3.0.war:/opt/guacamole/guacamole.war
> \
> > -e MYSQL_DATABASE=guacamole \
> > -e MYSQL_USER=root \
> > -e MYSQL_PASSWORD=guacamole \
> > -e GUACAMOLE_HOME=/etc/guacamole \
> > -d -p 8080:8080 \
> > guacamole/guacamole
> >
> > Building it from repo, yes.
> >
> > Manually layering an externally-built, newer .war file over the .war
> within the Docker image from a previous release is a bit funky. To ensure
> things are sane, can you build the Docker image directly from the same
> source and get rid of the volume mount? The Dockerfile for the image is
> part of the guacamole-client repository, so you can just do a "docker
> build" from the root of the source that you cloned from git to produce an
> updated image.
> >
> > Thanks,
> >
> > Michael Jumper
> > CEO, Lead Developer
> > Glyptodon Inc.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@guacamole.apache.org
> For additional commands, e-mail: user-help@guacamole.apache.org
>
>