Thank you for taking the time to review the issue. I'll try to ask some
information from an upstream point of view. Also, your remarks are
interesting wrt our policy of creating the final stable release.
I will start by the end: all our stable releases 2.x.y with y>0 are a
form of this cherry picking. We have a stable release manager who vets
such backporting on a case-by-case basis. We do not have comprehensive
regression tests for all these changes because (1) we are a small team
and (2) it is more difficult with a GUI.
Since I am on the issue of tests, what we have now is
* a few unit tests
* a reasonable set of tests for tex2lyx (importation from LaTeX)
* a reasonable set of tests of all possible export cases, although in this case
we mostly check that the exported LaTeX actually compiles. We do catch bugs
with it, though.
* we routinely run coverity scan on master branch and backport meaningful
fixes, and have a policy of having no compile warnings with -Wall -Wextra.
I understand that this is not professional-level quality insurance, but
at least we do try :)
The two big chunks of interest are:
* boost
- update included boost to 1.60: this is typically not used in a ubuntu
build, which uses system
library
- updates to compile with boost 1.60: this is needed in general since we want
LyX to be compilable on
more modern distributions.
I understand that these changes are not necessary for xenial.
* lyx2lyx: I am not sure what to think about that. I used to think that this
was definitely the way to
do it, but now I see your point about the amount of changes. All I can say is
that reading old files,
along with not losing work on a crash, are two things we are very careful
with. This is what it takes
to have users who trust us. Note that, thanks to lyx2lyx, LyX is still able
to read and typeset files
from the 0.7 era (1995).
I completely see your point about manual cherry-picking. However, we do
not have people doing that for each distribution. What I think is that
the work the we do at a global level with these stable distributions is
the best we can do given our lack of manpower. We try to maintain a real
difference between what goes into stable releases vs master branch.
What we want to setup soon is a service of nightly builds for at least
ubuntu so that at a few users get a chance to catch bugs earlier.
I will end with another argument towards your proto-decison of rejecting
the update: we do provide a ppa (lyx-devel) which provides recent lyx
builds for older distributions, under our responsibility. This is not
totally upd to date these days, as I have to take it over and I do not
understand anything about launchpad :)
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1595358
Title:
[SRU] Update to maintenance release 2.1.5 in Xenial
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/lyx/+bug/1595358/+subscriptions
--
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
