[Linux] Brackets depends on obsolete libgcrypt11 package which is no longer included by default

[Romane-T]

Brackets not currently installable on Jessie. Brackets has dependency on libgcrypt11, which is not available in the Jessie repositories. Jessie repositories have libgcrypt20. Brackets can be installed if user locates a copy of libgcrypt11, which can be co-installed with libgryypt20, but considering Jessie in freeze to become stable, doing this this should be unnecessary. Removing libgcrypt11 from system removes Brackets at the same time.
Romane

[prksingh]

@Romane-T Hey, could you try to build the appshell on your system.
The repo is located at: https://github.com/adobe/brackets-shell.
The build instructions are at: https://github.com/adobe/brackets-shell/wiki/Building-Brackets-Shell
This may help us locate if the error is with the installer.

You may want to check out this PR: adobe/brackets-shell#489
This may fix your issue but its not ready for testing yet.

[Romane-T]

On 21/01/15 21:17, Prashant Kumar Singh wrote:

@Romane-T https://github.com/Romane-T Hey, could you try to build
the appshell on your system.
The repo is located at: https://github.com/adobe/brackets-shell.
The build instructions are at:
https://github.com/adobe/brackets-shell/wiki/Building-Brackets-Shell
This may help us locate if the error is with the installer.

You may want to check out this PR: adobe/brackets-shell#489
adobe/brackets-shell#489
This may fix your issue but its not ready for testing yet.

—
Reply to this email directly or view it on GitHub
#10255 (comment).

Good morning

Many thanks for your reply.

Sorry, but looked at the material at the links given, and too much over
my head. The build instructions, though they seemed sort of straight
forward, still left me confused. The last link to the "PR" - well,
first, I have utterly no idea what a "PR" is, and what was on that page
made absolutely zero sense to me.

I am not a programmer. All that I learned about programming was a bit
more than a quarter a century ago. Everything has dramatically since
then, and I have forgotten it all anyway. I use Brackets solely to write
my HTML pages, and even then fairly simple.

I have no idea what the difference between libgcrypt11 and libgcrypt20
is, and have no interest in finding out. That is a developer issue.

Perhaps I am being simplistic, but it strikes me as a case of Brackets
declaring a dependency for libgcrypt11, while Debian have updated in
Jessie to libgcrypt20. If I remove libgcrypt11, Brackets is removed -
that's the nature of a dependency. Debian does not carry Brackets in its
repositories, so the issue returns to Brackets keeping up to what the
distribution does.

Reason and logic thus tells me that attempting to do the build you
suggested is unlikely to fix the problem UNLESS this updates the
dependencies for libgcrypt. If it doesn't update that dependency, then
the problem still exists, and attempting to remove libgcrypt11 will
still remove Brackets. And if it is the fix to the problem, then I am
certain that at least one of your developers who knows something more
about building and deploying will be quite capable of testing it and
applying it for the next release, which I would assume they are doing
anyway.

And while it is not fixed, Jessie users, except the small minority who
keep a copy of libgcrypt11 on hand, will not be users of Brackets,
noting that Jessie is soon to go stable (172 bugs left to squash).

With greetings

Romane

[prksingh]

I think libgcrypt11 is required for CEF(https://code.google.com/p/chromiumembedded/) that most brackets features run on. The link i had posted is a link to a pull request that upgrades to a more recent version of CEF (I think libcrypt20 will work with the new CEF)

Here is where you can download the build for linux: https://github.com/adobe/brackets/releases/tag/linux-cef-2171

I don't have access to a system with Debian so I can't be sure this build will fix your issue. There may be a few other small glitches as this build is still being tested. I will try to update the release with a more stable build once ready. Thanks!

[Romane-T]

On 23/01/15 21:44, Prashant Kumar Singh wrote:

I think libgcrypt11 is required for
CEF(https://code.google.com/p/chromiumembedded/) that most brackets
features run on. The link i had posted is a link to a pull request
that upgrades to a more recent version of CEF (I think libcrypt20 will
work with the new CEF)

Here is where you can download the build for linux:
https://github.com/adobe/brackets/releases/tag/linux-cef-2171

I don't have access to a system with Debian so I can't be sure this
build will fix your issue. There may be a few other small glitches as
this build is still being tested. I will try to update the release
with a more stable build once ready. Thanks!

—
Reply to this email directly or view it on GitHub
#10255 (comment).

Good morning

Now this is something within my capacity, and my thanks for giving me
the opportunity.

Your first paragraph, I gather that whatever a CEF is, the new version
might work with libgcrypt20

Downloaded the build in the second link, and before installing it,
purged the version already installed. Once purged, tested to see if
removing libgcrypt11 would affect any other package. There were no other
dependencies.

Installed the version just downloaded from the link in your email. The
install went without a hitch - zero warnings or errors.

Opened the application to ensure that it ran fine. it did. Noticed (an
aside to the topic of this thread) on closing that, though still needing
two clicks on the x button, the first click greyed the entire
application window, then on the second click the window closed as
expected. Previously, the first click on the x button did not appear to
do anything, so it would appear that some progress has been made to
resolving this issue. Closing through the 'File' menu closed the
application normally.

After closing Brackets, attempted to remove libgcrypt11. Unfortunately,
removing libgcrypt11 will still remove Brackets with it, so this build
(Release 1.2 experimental build 1.2.0-15628 (master 5dddaa1) build
timestamp: Thu Jan 22 2015 03:59:17 GMT-0800.) still retains the issue.

I even tried a reboot, knowing with 99% certainty that it would make no
difference. It didn't.

With warm greetings

Romane

[marcelgerber]

Some clarification here:
PR is short for "Pull Request" and it's when somebody proposes a change to the source code. It's pretty GitHub-specific, and as GitHub's all about social coding, it's a huge feature over here.

CEF is the "Chromium Embedded Framework", which is basically a configurable and embeddable version of the Chrome/Chromium browser. It's used in Brackets to render the UI, as Brackets is mostly written in HTML, CSS, and JavaScript - just like your website.

[Romane-T]

Good morning

This is, for a complete non-programmer as myself (I don't count the HTML
I do as programming), quite useful information. Not just for the
explanation of the acronyms - I feel a little less like an alien now in
a foreign environment - but for the clear fact, combining information in
both your and Prashant's posts, that the issue itself is not Brackets so
much as upstream. Till upstream provide an update that resolves the
issue, Brackets has its hands tied.

This is the only issue with Brackets and Jessie (Debian 8) that can
become a show-stopper. Not difficult to resolve if one knows where to
look. Have just checked to make sure, and Debian stable (Wheezy), due to
become old-stable when Jessie is declared stable, has libgcrypt11. It
will remain old-stable till the successor to Jessie goes stable and
Jessie becomes old-stable, which gives a fair window of time during
which the Chromium Embedded Framework can do the catchup. Perhaps just a
note in a relevant place for the unaware that this dependency exists for
Brackets at this time, and how they can obtain it - just a random
thought. Installing via dpkg is then a breeze.

In the interim, if you would like someone to provide any testing on
future development versions of Brackets to assess if this issue has been
resolved, feel free to ask. It will be my pleasure, even if it is only a
very tiny contribution back to the very excellent Brackets.

With greetings

Romane

On 24/01/15 00:11, Marcel Gerber wrote:

Some clarification here:
PR is short for "Pull Request" and it's when somebody proposes a
change to the source code. It's pretty GitHub-specific, and as
GitHub's all about social coding, it's a huge feature over here.

CEF is the "Chromium Embedded Framework", which is basically a
configurable and embeddable version of the Chrome/Chromium browser.
It's used in Brackets to render the UI, as Brackets is mostly written
in HTML, CSS, and JavaScript - just like your website.

—
Reply to this email directly or view it on GitHub
#10255 (comment).

[peterflynn]

Note: the issue with the close button is tracked as #4611, and it is indeed expected to be slightly improved in the test build linked above

[peterflynn]

I think to fully fix this we need to update the dependencies on our .deb build script to match the dependencies of the new CEF. @jasonsanjose said this is kind of a pain (manual process), but we've done it in the past for some earlier CEF updates, and it seems necessary this time.

[Romane-T]

Good morning

On 28/01/15 07:35, Peter Flynn wrote:

I think to fully fix this we need to update the dependencies on our
.deb build script to match the dependencies of the new CEF.
@jasonsanjose https://github.com/jasonsanjose said this is kind of a
pain (manual process), but we've done it in the past for some earlier
CEF updates, and it seems necessary this time.

—
Reply to this email directly or view it on GitHub
#10255 (comment).

Am I right in assuming that it is a tedious process rather than a
programming process? If not a programming process, for someone like
myself who would need perhaps some (or likely more) hand-holding as he

a) has no idea how to do a pull request and the associated return of
changed data
b) no idea of what actually needs to be changed

but is willing to learn provided there is someone with patience who can
answer sometimes (often?) silly questions (in other words, I can appear
really dumb sometimes), be of any use? Think of me as noob from core to
surface.

The doing will, in the final analysis, prove the best teacher.

Besides (Romane grinning), it would give me something to do every third
day. rather than my current cycle of every second day with existing
projects.

With greetings

Romane