Re: [blfs-dev] Tagging status

[Pierre Labastie]

On 26/02/2018 03:15, Bruce Dubbs wrote:
> Bruce Dubbs wrote:
>> Bruce Dubbs wrote:
>>> Bruce Dubbs wrote:
 Bruce Dubbs wrote:
> We are doing very well for the first weekend after package freeze:
>
> lfs81 tags remaining: 518
> lfs82 tags  : 274
>
> Tickets targeted for 8.2: 34

 ===

 As of Tue Feb 20 05:46:29 UTC 2018

 lfs81 tags remaining: 457
 lfs82 tags  : 335

 Tickets targeted for 8.2: 20

 I have a bunch of multimedia packages ready (~45) to tag, but not tonight.
>>>
>>> ===
>>>
>>> As of Thu Feb 22 05:07:06 UTC 2018
>>>
>>> lfs81 tags remaining: 272
>>> lfs82 tags  : 521
>>>
>>> Tickets targeted for 8.2: 16
>>>
>>> I tried to get the feh and mercurial tickets done today but am still trying
>>> to straighten out some issues.  Some of these packages require a LOT of
>>> dependencies and have substantial circular dependencies if trying to build
>>> all the optional deps.
>>
>> Sat Feb 24 22:42:59 CST 2018
>>
>> lfs81 tags remaining: 199
>> lfs82 tags  : 594
>>
>> Tickets targeted for 8.2: 13
>>
>> I finished up kf5 today and will get the kf5-apps and plasma tomorrow.
>> I also plan to do lxde and some other misc packages.
>>
>> A lot of the outstanding packages are systemd related, mostly gnome.
>>
>> DJ and I ran into some problems with java.  We can't build junit or
>> freetts.  We probably need to update juint to version 4.12.  Not sure about
>> freetts and I haven't checked fop or IcedTea-Web.
> 
> Sun Feb 25 20:06:54 CST 2018
> 
> lfs81 tags remaining: 99
> lfs82 tags  : 693
> 
> Tickets targeted for 8.2: 11
> 
> The junit problem has been solved.  We decided to archive freetts until next
> cycle and possibly look for another text-to-speech app.
> 
> New problem apps: juffed, abiword, balsa, alpine (url needs update but we have
> the tarball).

juffed works for me. What is the problem? If it is a problem with a library
not found, I've fixed it (bad link in qscintilla) (r19867).
Actually the right fix should be to use "readlink", as we do in LFS.

There is a ticket for alpine already (#10188).

For abiword, I'll try the patch, and let you know.

Not tested balsa (yet).

Pierre

-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] ptlib issue

[Pierre Labastie]

On 25/02/2018 23:50, Bruce Dubbs wrote:
> Pierre Labastie wrote:
>> On 25/02/2018 23:03, Bruce Dubbs wrote:
>>> Pierre Labastie wrote:
 On 25/02/2018 01:57, Bruce Dubbs wrote:
> There was a problem with ptlib and glibc-2.27.  I came up with a sed in
> addition to the other patches and seds:
>
> sed -e '/WCHAR;/s:^://:' \
>   -e '/UINT;/ s:^://:' \
>   -e '/limits.h/a #include sqltypes.h' \
>   -i include/ptlib/unix/ptlib/contain.h
>
> I would like to get someone to test it to confirm the it does allow the
> build
> to complete.
>>>
 I used a different solution: pass --disable-odbc to configure: the type
 conflict is with the unix-odbc headers. I do not think it comes from glibc,
 but it only shows up when unix-odbc is installed...
>>>
>>> Hmm.  Do we want to disable odbc?  It may not make any difference for us, 
>>> but
>>> the problem is a conflict between the code and changes to system libraries.
>>> Adding an include will not affect the code intended by upstream.
>>>
>>
>> I do not think there is a conflict with system libraries. It is a conflict
>> between ptlib libraries, and odbc libraries. As of disabling odbc (which is a
>> compatibility library for Winxx), nothing in the book uses ptlib with odbc
>> (and actually, odbc is deemed optional). So I do not think there is much harm
>> in disabling it.
> 
> OK, go ahead and make the change.  Be sure to mention the issue or at least
> say that the build is broken with it.
> 
>   -- Bruce
> 

Done at revision 19879

Pierre
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] Tagging status

[Bruce Dubbs]

Ken Moffat wrote:

On Sun, Feb 25, 2018 at 11:24:05PM +0100, Pierre Labastie wrote:


Also a few things I am seeking advice about: On the Mutt page, there is: "Mutt
will rebuild the html documentation if libxslt is present, then use that to
update the text manual - even if none of the specified browsers are present.
That will create an empty text file, so we will save the shipped file.", and
then, nothing is saved. Should this sentence be commented out?


I _thought_ I had noticed the omission of the copy when I updated to
1.9.3, and decided it was valid, but I've just now rebuilt and got
an empty text manual.  My memory says that I noticed it at a late
stage, and never got around to updating my own script.

Trying to identify in which commit something disappeared is not
easy, still looking - but it ought to be reinstated.

Hmm, based on a suggestion at stackoverflow, of the form
svn log --diff src/fozbo.cpp -r:HEAD

suggests _I_ took it out in r19714, but trac doesn't show that:
http://wiki.linuxfromscratch.org/blfs/changeset/19714


Looks like it was r19623.  You missed it by 1 commit.

http://wiki.linuxfromscratch.org/blfs/changeset?reponame==19623%40trunk%2FBOOK%2Fnetworking%2Fmailnews%2Fmutt.xml=19291%40trunk%2FBOOK%2Fnetworking%2Fmailnews%2Fmutt.xml

  -- Bruce
--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] Tagging status

[Ken Moffat]

On Sun, Feb 25, 2018 at 11:24:05PM +0100, Pierre Labastie wrote:
> 
> Also a few things I am seeking advice about: On the Mutt page, there is: "Mutt
> will rebuild the html documentation if libxslt is present, then use that to
> update the text manual - even if none of the specified browsers are present.
> That will create an empty text file, so we will save the shipped file.", and
> then, nothing is saved. Should this sentence be commented out?
> 
I _thought_ I had noticed the omission of the copy when I updated to
1.9.3, and decided it was valid, but I've just now rebuilt and got
an empty text manual.  My memory says that I noticed it at a late
stage, and never got around to updating my own script.

Trying to identify in which commit something disappeared is not
easy, still looking - but it ought to be reinstated.

Hmm, based on a suggestion at stackoverflow, of the form
svn log --diff src/fozbo.cpp -r:HEAD

suggests _I_ took it out in r19714, but trac doesn't show that:
http://wiki.linuxfromscratch.org/blfs/changeset/19714

Anyway, I'll reinstate it.

ĸen
-- 
Truth, in front of her huge walk-in wardrobe, selected black leather
boots with stiletto heels for such a barefaced truth.
 - Unseen Academicals
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] Tagging status

[Bruce Dubbs]

Bruce Dubbs wrote:

Bruce Dubbs wrote:

Bruce Dubbs wrote:

Bruce Dubbs wrote:

We are doing very well for the first weekend after package freeze:

lfs81 tags remaining: 518
lfs82 tags  : 274

Tickets targeted for 8.2: 34


===

As of Tue Feb 20 05:46:29 UTC 2018

lfs81 tags remaining: 457
lfs82 tags  : 335

Tickets targeted for 8.2: 20

I have a bunch of multimedia packages ready (~45) to tag, but not tonight.


===

As of Thu Feb 22 05:07:06 UTC 2018

lfs81 tags remaining: 272
lfs82 tags  : 521

Tickets targeted for 8.2: 16

I tried to get the feh and mercurial tickets done today but am still 
trying to straighten out some issues.  Some of these packages require a 
LOT of dependencies and have substantial circular dependencies if trying 
to build all the optional deps.


Sat Feb 24 22:42:59 CST 2018

lfs81 tags remaining: 199
lfs82 tags  : 594

Tickets targeted for 8.2: 13

I finished up kf5 today and will get the kf5-apps and plasma tomorrow.
I also plan to do lxde and some other misc packages.

A lot of the outstanding packages are systemd related, mostly gnome.

DJ and I ran into some problems with java.  We can't build junit or 
freetts.  We probably need to update juint to version 4.12.  Not sure 
about freetts and I haven't checked fop or IcedTea-Web.


Sun Feb 25 20:06:54 CST 2018

lfs81 tags remaining: 99
lfs82 tags  : 693

Tickets targeted for 8.2: 11

The junit problem has been solved.  We decided to archive freetts until 
next cycle and possibly look for another text-to-speech app.


New problem apps: juffed, abiword, balsa, alpine (url needs update but we 
have the tarball).


  -- Bruce


--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] missing alpine-2.21 download location

[Bruce Dubbs]

Alain Toussaint wrote:

for package alpine-2.21 
(http://linuxfromscratch.org/blfs/view/systemd/basicnet/alpine.html)

apparently, alpine is no longer available at the url referenced below:

Resolving alpine.freeiz.com (alpine.freeiz.com)... 145.14.144.8, 
2a02:4780:dead:f265::1
Connecting to alpine.freeiz.com (alpine.freeiz.com)|145.14.144.8|:80... 
connected.
HTTP request sent, awaiting response... 410 Gone
2018-02-25 18:14:13 ERROR 410: Gone.


checking up the gentoo ebuild, the right wget command should be:

wget 
http://repo.or.cz/alpine.git/snapshot/672d6838a9babf2faeb9f79267525a4ab9d20b14.tar.gz
 -O
alpine-2.21.1.tar.gz


Try https://ftp.osuosl.org/pub/blfs/conglomeration/alpine/alpine-2.21.tar.xz

  -- Bruce
--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

[blfs-dev] missing alpine-2.21 download location

[Alain Toussaint]

for package alpine-2.21 
(http://linuxfromscratch.org/blfs/view/systemd/basicnet/alpine.html)

apparently, alpine is no longer available at the url referenced below:

Resolving alpine.freeiz.com (alpine.freeiz.com)... 145.14.144.8, 
2a02:4780:dead:f265::1
Connecting to alpine.freeiz.com (alpine.freeiz.com)|145.14.144.8|:80... 
connected.
HTTP request sent, awaiting response... 410 Gone
2018-02-25 18:14:13 ERROR 410: Gone.


checking up the gentoo ebuild, the right wget command should be:

wget 
http://repo.or.cz/alpine.git/snapshot/672d6838a9babf2faeb9f79267525a4ab9d20b14.tar.gz
 -O
alpine-2.21.1.tar.gz

Alain
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] Tagging status

[Bruce Dubbs]

Pierre Labastie wrote:


Been there too. For junit, I have added a sed to the book, which allows to
build junit-4.11. For junit-4.12, or junit-5.0.x, the build system is maven,
so it is a big change, and not suitable for 8.2. Note that we should move to
maven for 8.3, since fop can be built with maven too (and is deprecating ant
builds).


I agree with that approach.  I tagged junit-4.11 earlier today.


For freetts, the same trick as for junit allows to compile, but then the
excutable fails with an exception about type casting. I'd suggest archiving
freetts now, and moving to flite after 8.2.


We can do that.  Please open a ticket for 8.3 though.  I'll archive freetts.


other issues:
I am also seeing the flickering in abiword, as reported in
https://bugzilla.abisource.com/show_bug.cgi?id=13791. Looks like the bug is
not fixed upstream. Do not know exactly what to do. The proposed patch is big,
and does not apply cleanly to recent versions of abiword.


I read through the bug.  Last entry was September.  Did you try changing 
themes?  abiword builds and runs for me.  I don't see any flickering, but 
I am running over ssh.  The cursor does seem to be a bit erratic as the 
blink rate varies and sometimes disappears completely for several seconds.


The only options I can see are to include a big warning or to just archive it.


Also a few things I am seeking advice about: On the Mutt page, there is: "Mutt
will rebuild the html documentation if libxslt is present, then use that to
update the text manual - even if none of the specified browsers are present.
That will create an empty text file, so we will save the shipped file.", and
then, nothing is saved. Should this sentence be commented out?


I haven't got to mutt yet, but I'd like to get Ken's opinion.


I do not like the sed in ptlib, I'd rather use --disable-odbc (nothing uses
ptlib with odbc anyway). See my preceding message about ptlib.


See my reply in the other thread.


And last, all the test in perl File::BaseDir in the userdirs.t part have
failed for me. I guess there is a conflict with xdg-user-dirs, which I am
using for LXQt. Should this be mentioned?


What does cpan -i File::BaseDir do?

  -- Bruce


--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] ptlib issue

[Bruce Dubbs]

Pierre Labastie wrote:

On 25/02/2018 23:03, Bruce Dubbs wrote:

Pierre Labastie wrote:

On 25/02/2018 01:57, Bruce Dubbs wrote:

There was a problem with ptlib and glibc-2.27.  I came up with a sed in
addition to the other patches and seds:

sed -e '/WCHAR;/s:^://:' \
  -e '/UINT;/ s:^://:' \
  -e '/limits.h/a #include sqltypes.h' \
  -i include/ptlib/unix/ptlib/contain.h

I would like to get someone to test it to confirm the it does allow the build
to complete.



I used a different solution: pass --disable-odbc to configure: the type
conflict is with the unix-odbc headers. I do not think it comes from glibc,
but it only shows up when unix-odbc is installed...


Hmm.  Do we want to disable odbc?  It may not make any difference for us, but
the problem is a conflict between the code and changes to system libraries.
Adding an include will not affect the code intended by upstream.



I do not think there is a conflict with system libraries. It is a conflict
between ptlib libraries, and odbc libraries. As of disabling odbc (which is a
compatibility library for Winxx), nothing in the book uses ptlib with odbc
(and actually, odbc is deemed optional). So I do not think there is much harm
in disabling it.


OK, go ahead and make the change.  Be sure to mention the issue or at 
least say that the build is broken with it.


  -- Bruce

--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] ptlib issue

[Pierre Labastie]

On 25/02/2018 23:03, Bruce Dubbs wrote:
> Pierre Labastie wrote:
>> On 25/02/2018 01:57, Bruce Dubbs wrote:
>>> There was a problem with ptlib and glibc-2.27.  I came up with a sed in
>>> addition to the other patches and seds:
>>>
>>> sed -e '/WCHAR;/s:^://:' \
>>>  -e '/UINT;/ s:^://:' \
>>>  -e '/limits.h/a #include sqltypes.h' \
>>>  -i include/ptlib/unix/ptlib/contain.h
>>>
>>> I would like to get someone to test it to confirm the it does allow the 
>>> build
>>> to complete.
> 
>> I used a different solution: pass --disable-odbc to configure: the type
>> conflict is with the unix-odbc headers. I do not think it comes from glibc,
>> but it only shows up when unix-odbc is installed...
> 
> Hmm.  Do we want to disable odbc?  It may not make any difference for us, but
> the problem is a conflict between the code and changes to system libraries. 
> Adding an include will not affect the code intended by upstream.
> 

I do not think there is a conflict with system libraries. It is a conflict
between ptlib libraries, and odbc libraries. As of disabling odbc (which is a
compatibility library for Winxx), nothing in the book uses ptlib with odbc
(and actually, odbc is deemed optional). So I do not think there is much harm
in disabling it.

Pierre

-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] vesa xorg-server driver missing

[Tim Tassonis]

On February 25, 2018 22:54:32 Pierre Labastie  wrote:



On 24/02/2018 18:45, Tim Tassonis wrote:
>

Hi all

>
>

I'm just about to build the xorg-server drivers and noticed that the vesa
driver has gone from the book. I'm not an expert in xorg and I could
successfully start my xserver using the fbdev driver in my qemu vm that I
started with vga=std. However, the current vesa driver
xf86-video-vesa-2.4.0.tar.gz builds and runs fine with current lfs/blfs.

>
>

Is there a reason that it got dropped, maybe like there be no actual hardware
or virtual environment requring it?




The qemu VM's with vga=std work pretty well with the KMS driver, which is much
better than the vesa one. ISTR that was the reason for removing the veas 
driver.


Yes, the x server actually unloads the vesa driver in that case again after 
loading it, so it is not even used when present. Guess, I'll droo it, too 
in my setup.


Bye
Tim




Pierre
--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page




--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] Tagging status

[Pierre Labastie]

On 25/02/2018 05:56, Bruce Dubbs wrote:
> Bruce Dubbs wrote:
>> Bruce Dubbs wrote:
>>> Bruce Dubbs wrote:
 We are doing very well for the first weekend after package freeze:

 lfs81 tags remaining: 518
 lfs82 tags  : 274

 Tickets targeted for 8.2: 34
>>>
>>> ===
>>>
>>> As of Tue Feb 20 05:46:29 UTC 2018
>>>
>>> lfs81 tags remaining: 457
>>> lfs82 tags  : 335
>>>
>>> Tickets targeted for 8.2: 20
>>>
>>> I have a bunch of multimedia packages ready (~45) to tag, but not tonight.
>>
>> ===
>>
>> As of Thu Feb 22 05:07:06 UTC 2018
>>
>> lfs81 tags remaining: 272
>> lfs82 tags  : 521
>>
>> Tickets targeted for 8.2: 16
>>
>> I tried to get the feh and mercurial tickets done today but am still trying
>> to straighten out some issues.  Some of these packages require a LOT of
>> dependencies and have substantial circular dependencies if trying to build
>> all the optional deps.
> 
> Sat Feb 24 22:42:59 CST 2018
> 
> lfs81 tags remaining: 199
> lfs82 tags  : 594
> 
> Tickets targeted for 8.2: 13
> 
> I finished up kf5 today and will get the kf5-apps and plasma tomorrow.
> I also plan to do lxde and some other misc packages.
> 
> A lot of the outstanding packages are systemd related, mostly gnome.
> 
> DJ and I ran into some problems with java.  We can't build junit or freetts. 
> We probably need to update juint to version 4.12.  Not sure about freetts and
> I haven't checked fop or IcedTea-Web.
> 

Been there too. For junit, I have added a sed to the book, which allows to
build junit-4.11. For junit-4.12, or junit-5.0.x, the build system is maven,
so it is a big change, and not suitable for 8.2. Note that we should move to
maven for 8.3, since fop can be built with maven too (and is deprecating ant
builds).

For freetts, the same trick as for junit allows to compile, but then the
excutable fails with an exception about type casting. I'd suggest archiving
freetts now, and moving to flite after 8.2.

other issues:
I am also seeing the flickering in abiword, as reported in
https://bugzilla.abisource.com/show_bug.cgi?id=13791. Looks like the bug is
not fixed upstream. Do not know exactly what to do. The proposed patch is big,
and does not apply cleanly to recent versions of abiword.

Also a few things I am seeking advice about: On the Mutt page, there is: "Mutt
will rebuild the html documentation if libxslt is present, then use that to
update the text manual - even if none of the specified browsers are present.
That will create an empty text file, so we will save the shipped file.", and
then, nothing is saved. Should this sentence be commented out?

I do not like the sed in ptlib, I'd rather use --disable-odbc (nothing uses
ptlib with odbc anyway). See my preceding message about ptlib.

And last, all the test in perl File::BaseDir in the userdirs.t part have
failed for me. I guess there is a conflict with xdg-user-dirs, which I am
using for LXQt. Should this be mentioned?

Pierre

-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] Report on my issue with gcc include next (missing math.h header)

[Alain Toussaint]

Le dimanche 25 février 2018 à 17:12 -0500, Alain Toussaint a écrit :
> > When does it fail when building with debian? Immediately in chapter 5?, or 
> > later?
> 
> 
> Much later, picture that workflow:
> 
> debian 9.3 --> alfs start a build --> boot the LFS + some BLFS (prerequisites 
> for another round of
> alfs) just built --> download alfs by svn, start another build (LFS only at 
> that point) which
> fail.
> 
> No optimization used in any of the builds and all final testsuites enabled; 
> one cpu core used for
> all make invocations.
> 
> Alain


forgot to add: at the last alfs step, the build break at chapter 5 pass 1.

Alain
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] Report on my issue with gcc include next (missing math.h header)

[Alain Toussaint]

> When does it fail when building with debian? Immediately in chapter 5?, or 
> later?


Much later, picture that workflow:

debian 9.3 --> alfs start a build --> boot the LFS + some BLFS (prerequisites 
for another round of
alfs) just built --> download alfs by svn, start another build (LFS only at 
that point) which fail.

No optimization used in any of the builds and all final testsuites enabled; one 
cpu core used for
all make invocations.

Alain
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] Report on my issue with gcc include next (missing math.h header)

[Pierre Labastie]

On 25/02/2018 04:15, Alain Toussaint wrote:
> 
>> Ah, so this is all specific to systems where gcc has been rebuilt to
>> add extra languages,
> 
> maybe, maybe not, which is why I'm trying to reproduce it under a brand new 
> user account with as few
> variables defined as possible under which, I downloaded ALFS trunk via svn 
> (ditto with LFS and
> BLFS).
> 
>>  and then it gave problems with:
>>
>> · building mesa
>>
>> · building a new LFS
> 
> 
> That's right. mesa was attempted once but the new LFS was retried many times 
> under different
> conditions (one LFS built by hand with the only change was to install BLFS 
> gcc with extra languages,
> another attempt under LFS which was built by ALFS on a debian base) and then 
> some more.
> 
> 
>> If so, maybe people who rebuild gcc to add extra
>> languages have some suggestions ?
> 
> Maybe but on this list, there has been no recent report since I've been 
> registered to this list (~2
> months ago). I'll go as far as I can and will ask for backup if needed (most 
> of my work positions
> have been like that. I take LFS as a job too).
> 
> Al
> 

I've been the maintainer of BLFS GCC for a while, and I haven't seen problems
with include_next, except, as I said, when headers "fixed" by fixinclude would
produce version mismatches.

When does it fail when building with debian? Immediately in chapter 5?, or 
later?

Pierre
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] ptlib issue

[Bruce Dubbs]

Pierre Labastie wrote:

On 25/02/2018 01:57, Bruce Dubbs wrote:

There was a problem with ptlib and glibc-2.27.  I came up with a sed in
addition to the other patches and seds:

sed -e '/WCHAR;/s:^://:' \
     -e '/UINT;/ s:^://:' \
     -e '/limits.h/a #include sqltypes.h' \
     -i include/ptlib/unix/ptlib/contain.h

I would like to get someone to test it to confirm the it does allow the build
to complete.



I used a different solution: pass --disable-odbc to configure: the type
conflict is with the unix-odbc headers. I do not think it comes from glibc,
but it only shows up when unix-odbc is installed...


Hmm.  Do we want to disable odbc?  It may not make any difference for us, 
but the problem is a conflict between the code and changes to system 
libraries.  Adding an include will not affect the code intended by upstream.


  -- Bruce

--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] ptlib issue

[Pierre Labastie]

On 25/02/2018 01:57, Bruce Dubbs wrote:
> There was a problem with ptlib and glibc-2.27.  I came up with a sed in
> addition to the other patches and seds:
> 
> sed -e '/WCHAR;/s:^://:' \
>     -e '/UINT;/ s:^://:' \
>     -e '/limits.h/a #include sqltypes.h' \
>     -i include/ptlib/unix/ptlib/contain.h
> 
> I would like to get someone to test it to confirm the it does allow the build
> to complete.
> 
>   -- Bruce

I used a different solution: pass --disable-odbc to configure: the type
conflict is with the unix-odbc headers. I do not think it comes from glibc,
but it only shows up when unix-odbc is installed...

Pierre


-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Re: [blfs-dev] vesa xorg-server driver missing

[Pierre Labastie]

On 24/02/2018 18:45, Tim Tassonis wrote:
> Hi all
> 
> I'm just about to build the xorg-server drivers and noticed that the vesa
> driver has gone from the book. I'm not an expert in xorg and I could
> successfully start my xserver using the fbdev driver in my qemu vm that I
> started with vga=std. However, the current vesa driver
> xf86-video-vesa-2.4.0.tar.gz builds and runs fine with current lfs/blfs.
> 
> Is there a reason that it got dropped, maybe like there be no actual hardware
> or virtual environment requring it?
> 
> 

The qemu VM's with vga=std work pretty well with the KMS driver, which is much
better than the vesa one. ISTR that was the reason for removing the veas driver.

Pierre
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page