Bug#890204: transition: ppp

2018-02-11 Thread Chris Boot 
On 11/02/18 21:04, Chris Boot wrote:
> Note that connman seems to be a package to newly depend on ppp-dev and
> builds a pppd plugin but it doesn't use my dh_ppp helper. I will be
> filing a bug against that package (including a patch) shortly. Even
> without said patch a binNMU is sufficient (and required) for binary
> compatibility with the new ppp upload.

Please see bug #890207 and the Salsa Merge Request at:
https://salsa.debian.org/debian/connman/merge_requests/1

Thanks,
Chris

-- 
Chris Boot
bo...@debian.org



signature.asc
Description: OpenPGP digital signature

Bug#890204: transition: ppp

2018-02-11 Thread Chris Boot 
Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: transition

Hi release team folks,

I have just uploaded a new version of ppp to experimental. That upload
requires packages that build pppd plugins to be rebuilt due to a change
in the pppd ABI. I would like to upload that package to sid as soon as
feasible.

This is the first opportunity we've had to try my new plugin ABI
tracking technique. I've given the ben file below a try myself and I
think it does the right thing, but I would appreciate any comments you
might have about it.

I have manually rebuilt all affected packages and all build fine so
binNMUs should be the only thing required for all of them.

Note that connman seems to be a package to newly depend on ppp-dev and
builds a pppd plugin but it doesn't use my dh_ppp helper. I will be
filing a bug against that package (including a patch) shortly. Even
without said patch a binNMU is sufficient (and required) for binary
compatibility with the new ppp upload.

Ben file:

title = "ppp";
is_affected = .build-depends ~ /ppp-dev/;
is_good = .depends ~ /ppp \(<< 2.4.7-3~\)/ | .breaks ~ /ppp \(>= 2.4.7-3~\)/;
is_bad = .depends ~ /ppp \(<< 2.4.7-2~\)/ | .breaks ~ /ppp \(>= 2.4.7-2~\)/;

Bug#884711: stretch-pu: package dpdk/16.11.4-1+deb9u1

2018-02-11 Thread Luca Boccassi 
On Sat, 2018-02-10 at 09:47 +0100, Julien Cristau wrote:
> Control: tags -1 confirmed
> 
> On Sun, Jan 14, 2018 at 20:14:14 +, Luca Boccassi wrote:
> 
> > Thank you for the review, I have reworked the debdiff - rather than
> > taking the last 16.11.x version we uploaded in Sid and reverting a
> > few
> > changes, instead I've taken the version in Stretch and merged the
> > LTS
> > point releases and a couple of small changes, details below.
> > 
> 
> Great, thanks.  Go ahead and upload to stretch.
> 
> Cheers,
> Julien

Done, thanks.

There will be one LTS bugfix release every 3 months until at least
November. I think there is no whitelisting system like Ubuntu has, and
so I will need to create a release.debian.org bug each time, right?

-- 
Kind regards,
Luca Boccassi

signature.asc
Description: This is a digitally signed message part

Bug#890189: stretch-pu: package libreoffice/1:5.2.7-1+deb9u3

2018-02-11 Thread Rene Engelhard 
Package: release.debian.org
Severity: normal
Tags: stretch
User: release.debian@packages.debian.org
Usertags: pu

This is a follow-up to the 1:5.2.7-1+deb9u2 security update.

- As was correctly pointed out on IRC the debdiff contains cruft from the
  1:5.2.7-1 backport to jessie:

--- snip ---
Base version: libreoffice_5.2.7-1
Target version: libreoffice_5.2.7-1+deb9u1
Base file: 
/srv/ftp-master.debian.org/ftp/pool/main/libr/libreoffice/libreoffice_5.2.7-1.dsc
Target file: 
/srv/ftp-master.debian.org/policy/pool/main/libr/libreoffice/libreoffice_5.2.7-1+deb9u1.dsc

[...]
 libreoffice-dbg.postinst.debhelper  |3 
 libreoffice-dbg.postrm.debhelper|3 
 libreoffice-dbg.preinst.debhelper   |3 
 libreoffice-dbg.prerm.debhelper |3 
 libreoffice-dbg.substvars   |4 
[...]
 uno-libs3-dbg.substvars |3 
 ure-dbg.substvars   |3 
--- snip ---

 They have been removed.

- Upstream pointed out that my backport of the 5.4 patches contain a
  minor bug flagging more stuff as an error than needed. If we are not
  mistaken this isn't as critical to warrant a new security round again,
  so... Fix:

@@ -445,7 +445,7 @@ index 0d55e87..b475b1c 100644
  
 -sal_Int32 nRead = 0;
 -while ( ( nRead = xStream->readBytes( buffer, BUF_LEN ) ) == BUF_LEN )
-+bool bWasError = (pMyFormulaCell && pMyFormulaCell->GetRawError() != 
formula::errNoCode);
++bool bWasError = (pMyFormulaCell && pMyFormulaCell->GetRawError() != 
formula::errNONE);
 +
 +if (!pLink)
  {
@@ -484,8 +484,8 @@ index 0d55e87..b475b1c 100644
  
 -xStream->closeInput();
 +//  If an new Error from Reschedule appears when the link is executed 
then reset the errorflag
-+if (pMyFormulaCell && pMyFormulaCell->GetRawError() != 
formula::errNoCode && !bWasError)
-+pMyFormulaCell->SetErrCode(formula::errNoCode);
++if (pMyFormulaCell && pMyFormulaCell->GetRawError() != 
formula::errNONE && !bWasError)
++pMyFormulaCell->SetErrCode(formula::errNONE);
 +
 +//  check the value
 +if (pLink->HasResult())
@@ -1104,3 +1104,15 @@ index 9de5851..586511d 100644
 -- 
 1.9.1
 
+diff --git a/include/formula/errorcodes.hxx b/include/formula/errorcodes.hxx
+index 173c744..5a5a8fd 100644
+--- a/include/formula/errorcodes.hxx
 b/include/formula/errorcodes.hxx
+@@ -24,6 +24,7 @@
+ 
+ namespace formula {
+ 
++const sal_uInt16 errNONE = 0;
+ const sal_uInt16 errIllegalChar  = 501;
+ const sal_uInt16 errIllegalArgument  = 502;
+ const sal_uInt16 errIllegalFPOperation   = 503; // #NUM!

 and changing two FormulaError::NoValue to the correct formula::errNoValue.
 (just for clarity, these are incremental patches, so this one not
 actually used)

- dk.mk is in -dev-common but is actually arch-specific. So it's moved
  to -dev. This file is needed for libreoffice-dev usage via the SDKs
  own mechanism.

- disabling tests on !amd64/!i386 (where it's notfatal) because the
  the jpeg filter (and thus vcl_filters_test) leaks memory and the test
  tries to use 100g on arm64/ppc64el builds and thus it got killed by
  jcristau. Fixed upstream but too big and intrusive to backport imho.

  -security doesn't want that fix in -security, so here it is.

Full debdiff to deb9u2 attached.

Regards,

Rene

-- System Information:
Debian Release: buster/sid
  APT prefers testing
  APT policy: (990, 'testing')
Architecture: arm64 (aarch64)

Kernel: Linux 4.14.0-3-arm64 (SMP w/4 CPU cores)
Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8), 
LANGUAGE=de_DE.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled
diff -Nru libreoffice-5.2.7/debian/changelog libreoffice-5.2.7/debian/changelog
--- libreoffice-5.2.7/debian/changelog  2018-02-09 07:22:22.0 +0100
+++ libreoffice-5.2.7/debian/changelog  2018-02-09 16:24:58.0 +0100
@@ -1,3 +1,16 @@
+libreoffice (1:5.2.7-1+deb9u3) stretch; urgency=medium
+
+  * debian/patches/WEBSERVICE-DDE.diff: improve to not throw more errors than
+neccessary (use the right error code) on WEBSERVICE() failures, thanks
+Jan-Marek Glogowski; do another 
s/FormulaError::NoValue/formula::errNoValue/
+
+  * debian/rules:
+- do not run the tests except on i386 (notfatal) and amd64
+- move dk.mk from -dev-common to -dev as it's not arch-indep, thanks
+  Rico Tzschichholz
+
+ -- Rene Engelhard   Fri, 09 Feb 2018 16:24:58 +0100
+
 libreoffice (1:5.2.7-1+deb9u2) stretch-security; urgency=high
 
   * fix control
diff -Nru libreoffice-5.2.7/debian/control libreoffice-5.2.7/debian/control
--- libreoffice-5.2.7/debian/control2018-02-09 07:21:38.0 +0100
+++ libreoffice-5.2.7/debian/control2018-02-09 16:24:58.0 +0100
@@ -4376,9 +4376,9 @@
 Suggests: libmythes-dev, libreoffice-dev-doc,

NEW changes in stable-new

2018-02-11 Thread Debian FTP Masters 
Processing changes file: tzdata_2018c-0+deb9u1_all.changes
  ACCEPT

Re: Scheduling 9.4

2018-02-11 Thread Cyril Brulebois 
Julien Cristau  (2018-02-10):
> we shipped 9.3 a couple of months ago, so we're overdue for 9.4.
> 
> Can you please let us know your availability on the following:
> - March 3
> - March 10
> - March 17
> - March 24
> - March 31

All dates equally OKish for me.

AFAICT:
 - There's a busybox I need to look at.
 - The kernel ABI bump is already in git, and we'll need a source upload
   accordingly.

debian-boot@ people, feel free to mention other things that should be in
the next stretch point release (ideally cc-ing me).


Cheers,
-- 
Cyril Brulebois (k...@debian.org)
D-I release manager -- Release team member -- Freelance Consultant


signature.asc
Description: PGP signature

Re: wanna-build access for binNMUs

2018-02-11 Thread Michael Stapelberg 
On Sat, Feb 10, 2018 at 1:14 PM, Philipp Kern  wrote:

> On 02/10/2018 09:26 AM, Julien Cristau wrote:
> > On Tue, Jan 30, 2018 at 20:45:44 +0100, Emilio Pozuelo Monfort wrote:
> >
> >> On 30/01/18 08:54, Michael Stapelberg wrote:
> >>> This would also be very helpful for fixing security issue #888777.
> >>
> >> You need to talk to the wanna-build team if you want to be able to
> schedule
> >> binNMUs for your language rebuilds, just like the ocaml and haskell
> teams do.
> >>
> > FWIW I disagree, I think this is something the wanna-build team has
> > essentially delegated to release, so getting people on board is a shared
> > thing between those two teams rather than solely a w-b thing.  It
> > happens once every few years though so there isn't really a process to
> > vet people.
>
> So can we have an opinion on the current proposal, which is a single DD
> (stapelberg)? As far as I know DSA's ticket is currently blocked on this.
>

nit: there’s no blocked ticket, but rather, once debian-release is okay
with the proposal, someone would need to file a ticket on my behalf. See my
other reply upthread:

> Hector Oron pointed out in the Debian RT that the release-team should
give their okay before the PostgreSQL account can be created. This should
be done by way of the release team filing the Debian RT ticket on my
behalf, preferable as an inline signed email.

If I understand https://www.debian.org/intro/organization correctly,
then pochu@ should be able to make this call. Given that no objections have
been voiced by anyone involved within the last two weeks since I asked for
access, may I suggest to go ahead? Thank you!


>
> Kind regards
> Philipp Kern
>
>
>


-- 
Best regards,
Michael

NEW changes in oldstable-new

2018-02-11 Thread Debian FTP Masters 
Processing changes file: libdatetime-timezone-perl_1.75-2+2018b_amd64.changes
  ACCEPT
Processing changes file: tzdata_2018c-0+deb8u1_amd64.changes
  ACCEPT

NEW changes in stable-new

2018-02-11 Thread Debian FTP Masters 
Processing changes file: libdatetime-timezone-perl_2.09-1+2018b_amd64.changes
  ACCEPT
Processing changes file: tzdata_2018c-0+deb9u1_source.changes
  ACCEPT

Bug#890125: jessie-pu: package uwsgi/2.0.7-1+deb8u2

2018-02-11 Thread Salvatore Bonaccorso 
Package: release.debian.org
Severity: normal
Tags: jessie
User: release.debian@packages.debian.org
Usertags: pu

Hi

The attached debdiff adresses CVE-2018-6758 / 889753 for uwsgi in
jessie. Cf.
http://lists.unbit.it/pipermail/uwsgi/2018-February/008835.html . The
issue does not warrant a DSA.

+uwsgi (2.0.7-1+deb8u2) jessie; urgency=medium
+
+  * Non-maintainer upload.
+  * Stack-based buffer overflow in uwsgi_expand_path function (CVE-2018-6758)
+(Closes: #889753)

Can it be considered for the upcoming jessie point release?

Regards,
Salvatore
diff -Nru uwsgi-2.0.7/debian/changelog uwsgi-2.0.7/debian/changelog
--- uwsgi-2.0.7/debian/changelog2017-02-19 22:17:18.0 +0100
+++ uwsgi-2.0.7/debian/changelog2018-02-06 21:39:42.0 +0100
@@ -1,3 +1,11 @@
+uwsgi (2.0.7-1+deb8u2) jessie; urgency=medium
+
+  * Non-maintainer upload.
+  * Stack-based buffer overflow in uwsgi_expand_path function (CVE-2018-6758)
+(Closes: #889753)
+
+ -- Salvatore Bonaccorso   Tue, 06 Feb 2018 21:39:42 +0100
+
 uwsgi (2.0.7-1+deb8u1) jessie; urgency=medium
 
   * Add patch cherry-picked upstream to fix compilation with recent
diff -Nru 
uwsgi-2.0.7/debian/patches/0001-improve-uwsgi_expand_path-to-sanitize-input-avoiding.patch
 
uwsgi-2.0.7/debian/patches/0001-improve-uwsgi_expand_path-to-sanitize-input-avoiding.patch
--- 
uwsgi-2.0.7/debian/patches/0001-improve-uwsgi_expand_path-to-sanitize-input-avoiding.patch
  1970-01-01 01:00:00.0 +0100
+++ 
uwsgi-2.0.7/debian/patches/0001-improve-uwsgi_expand_path-to-sanitize-input-avoiding.patch
  2018-02-06 21:39:42.0 +0100
@@ -0,0 +1,46 @@
+From: Unbit 
+Date: Tue, 6 Feb 2018 16:01:47 +0100
+Subject: improve uwsgi_expand_path() to sanitize input, avoiding stack
+ corruption and potential security issue
+Origin: 
https://github.com/unbit/uwsgi/commit/cb4636f7c0af2e97a4eef7a3cdcbd85a71247bfe
+Bug-Debian: https://bugs.debian.org/889753
+Bug-Debian-Security: https://security-tracker.debian.org/tracker/CVE-2018-6758
+
+---
+ core/utils.c | 11 ---
+ 1 file changed, 8 insertions(+), 3 deletions(-)
+
+diff --git a/core/utils.c b/core/utils.c
+index b4c98dfd..fd886932 100644
+--- a/core/utils.c
 b/core/utils.c
+@@ -3674,9 +3674,12 @@ void uwsgi_write_pidfile_explicit(char *pidfile_name, 
pid_t pid) {
+ }
+ 
+ char *uwsgi_expand_path(char *dir, int dir_len, char *ptr) {
+-  char src[PATH_MAX + 1];
+-  memcpy(src, dir, dir_len);
+-  src[dir_len] = 0;
++  if (dir_len > PATH_MAX)
++  {
++  uwsgi_log("invalid path size: %d (max %d)\n", dir_len, 
PATH_MAX);
++  return NULL;
++  }
++  char *src = uwsgi_concat2n(dir, dir_len, "", 0);
+   char *dst = ptr;
+   if (!dst)
+   dst = uwsgi_malloc(PATH_MAX + 1);
+@@ -3684,8 +3687,10 @@ char *uwsgi_expand_path(char *dir, int dir_len, char 
*ptr) {
+   uwsgi_error_realpath(src);
+   if (!ptr)
+   free(dst);
++  free(src);
+   return NULL;
+   }
++  free(src);
+   return dst;
+ }
+ 
+-- 
+2.11.0
+
diff -Nru uwsgi-2.0.7/debian/patches/series uwsgi-2.0.7/debian/patches/series
--- uwsgi-2.0.7/debian/patches/series   2017-02-08 20:11:59.0 +0100
+++ uwsgi-2.0.7/debian/patches/series   2018-02-06 21:39:42.0 +0100
@@ -1,4 +1,5 @@
 020150415~f6e5db9.patch
+0001-improve-uwsgi_expand_path-to-sanitize-input-avoiding.patch
 1001_avoid_setting_RPATH.patch
 1002_fix-reload-process-name.patch
 1003_remove-php-libs.patch

Bug#890105: stretch-pu: package uwsgi/2.0.14+20161117-3+deb9u1

2018-02-11 Thread Salvatore Bonaccorso 
Package: release.debian.org
Severity: normal
Tags: stretch
User: release.debian@packages.debian.org
Usertags: pu

Hi

The attached debdiff adresses CVE-2018-6758 / 889753 for uwsgi in
stretch. Cf.
http://lists.unbit.it/pipermail/uwsgi/2018-February/008835.html . The
issue does not warrant a DSA.

+uwsgi (2.0.14+20161117-3+deb9u1) stretch; urgency=medium
+
+  * Non-maintainer upload.
+  * Stack-based buffer overflow in uwsgi_expand_path function (CVE-2018-6758)
+(Closes: #889753)

Can it be considered for the upcoming stretch point release?

Regards,
Salvatore
diff -Nru uwsgi-2.0.14+20161117/debian/changelog 
uwsgi-2.0.14+20161117/debian/changelog
--- uwsgi-2.0.14+20161117/debian/changelog  2016-12-07 17:01:16.0 
+0100
+++ uwsgi-2.0.14+20161117/debian/changelog  2018-02-06 22:31:14.0 
+0100
@@ -1,3 +1,11 @@
+uwsgi (2.0.14+20161117-3+deb9u1) stretch; urgency=medium
+
+  * Non-maintainer upload.
+  * Stack-based buffer overflow in uwsgi_expand_path function (CVE-2018-6758)
+(Closes: #889753)
+
+ -- Salvatore Bonaccorso   Tue, 06 Feb 2018 22:31:14 +0100
+
 uwsgi (2.0.14+20161117-3) unstable; urgency=medium
 
   * Add patch 1007 to fix build router_basicauth for GNU/kFreeBSD.
diff -Nru 
uwsgi-2.0.14+20161117/debian/patches/0001-improve-uwsgi_expand_path-to-sanitize-input-avoiding.patch
 
uwsgi-2.0.14+20161117/debian/patches/0001-improve-uwsgi_expand_path-to-sanitize-input-avoiding.patch
--- 
uwsgi-2.0.14+20161117/debian/patches/0001-improve-uwsgi_expand_path-to-sanitize-input-avoiding.patch
1970-01-01 01:00:00.0 +0100
+++ 
uwsgi-2.0.14+20161117/debian/patches/0001-improve-uwsgi_expand_path-to-sanitize-input-avoiding.patch
2018-02-06 22:31:14.0 +0100
@@ -0,0 +1,46 @@
+From: Unbit 
+Date: Tue, 6 Feb 2018 16:01:47 +0100
+Subject: improve uwsgi_expand_path() to sanitize input, avoiding stack
+ corruption and potential security issue
+Origin: 
https://github.com/unbit/uwsgi/commit/cb4636f7c0af2e97a4eef7a3cdcbd85a71247bfe
+Bug-Debian: https://bugs.debian.org/889753
+Bug-Debian-Security: https://security-tracker.debian.org/tracker/CVE-2018-6758
+
+---
+ core/utils.c | 11 ---
+ 1 file changed, 8 insertions(+), 3 deletions(-)
+
+diff --git a/core/utils.c b/core/utils.c
+index b4c98dfd..fd886932 100644
+--- a/core/utils.c
 b/core/utils.c
+@@ -3674,9 +3674,12 @@ void uwsgi_write_pidfile_explicit(char *pidfile_name, 
pid_t pid) {
+ }
+ 
+ char *uwsgi_expand_path(char *dir, int dir_len, char *ptr) {
+-  char src[PATH_MAX + 1];
+-  memcpy(src, dir, dir_len);
+-  src[dir_len] = 0;
++  if (dir_len > PATH_MAX)
++  {
++  uwsgi_log("invalid path size: %d (max %d)\n", dir_len, 
PATH_MAX);
++  return NULL;
++  }
++  char *src = uwsgi_concat2n(dir, dir_len, "", 0);
+   char *dst = ptr;
+   if (!dst)
+   dst = uwsgi_malloc(PATH_MAX + 1);
+@@ -3684,8 +3687,10 @@ char *uwsgi_expand_path(char *dir, int dir_len, char 
*ptr) {
+   uwsgi_error_realpath(src);
+   if (!ptr)
+   free(dst);
++  free(src);
+   return NULL;
+   }
++  free(src);
+   return dst;
+ }
+ 
+-- 
+2.11.0
+
diff -Nru uwsgi-2.0.14+20161117/debian/patches/series 
uwsgi-2.0.14+20161117/debian/patches/series
--- uwsgi-2.0.14+20161117/debian/patches/series 2016-12-07 16:24:30.0 
+0100
+++ uwsgi-2.0.14+20161117/debian/patches/series 2018-02-06 22:31:14.0 
+0100
@@ -1,4 +1,5 @@
 020161123~6b48ac0.patch
+0001-improve-uwsgi_expand_path-to-sanitize-input-avoiding.patch
 1001_avoid_setting_RPATH.patch
 1002_fix-reload-process-name.patch
 1003_remove-php-libs.patch

Bug#885582: stretch-pu: package ncurses/6.0+20161126-1+deb9u2

2018-02-11 Thread Sven Joachim 
On 2018-02-11 09:45 +0100, Julien Cristau wrote:

> Control: tag -1 - moreinfo
> Control: tag -1 confirmed
>
> On Sat, Feb 10, 2018 at 11:08:37 +0100, Julien Cristau wrote:
>
>> Control: tag -1 moreinfo
>> 
> Got that one wrong, sorry.

Thanks, uploaded.

Cheers,
   Sven

Bug#890101: transition: libopenraw

2018-02-11 Thread Matthias Klose 
Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: transition

libopenraw is now uploaded to unstable.  The only package needing a binNMU is
tumbler.

Processed: Re: Bug#885582: stretch-pu: package ncurses/6.0+20161126-1+deb9u2

2018-02-11 Thread Debian Bug Tracking System 
Processing control commands:

> tag -1 - moreinfo
Bug #885582 [release.debian.org] stretch-pu: package 
ncurses/6.0+20161126-1+deb9u2
Removed tag(s) moreinfo.
> tag -1 confirmed
Bug #885582 [release.debian.org] stretch-pu: package 
ncurses/6.0+20161126-1+deb9u2
Added tag(s) confirmed.

-- 
885582: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=885582
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems

Bug#885582: stretch-pu: package ncurses/6.0+20161126-1+deb9u2

2018-02-11 Thread Julien Cristau 
Control: tag -1 - moreinfo
Control: tag -1 confirmed

On Sat, Feb 10, 2018 at 11:08:37 +0100, Julien Cristau wrote:

> Control: tag -1 moreinfo
> 
Got that one wrong, sorry.

Cheers,
Julien

Bug#885582: stretch-pu: package ncurses/6.0+20161126-1+deb9u2

2018-02-11 Thread Sven Joachim 
On 2018-02-10 11:08 +0100, Julien Cristau wrote:

> Control: tag -1 moreinfo
> [...]
> Thanks, go ahead.

This is contradictory.  Did you meant to tag the bug "confirmed" rather
than "moreinfo"?

>> +--- a/ncurses/tinfo/write_entry.c
>>  b/ncurses/tinfo/write_entry.c
>> +@@ -267,6 +267,9 @@ _nc_write_entry(TERMTYPE *const tp)
>> + #endif
>> + #endif /* USE_SYMLINKS */
>> + 
>> ++unsigned limit2 = sizeof(filename) - (2 + LEAF_LEN);
>> ++char saved = '\0';
>> ++
>> + static int call_count;
>> + static time_t start_time;  /* time at start of writes */
>> + 
>> +@@ -365,12 +368,18 @@ _nc_write_entry(TERMTYPE *const tp)
>> +start_time = 0;
>> + }
>> + 
>> +-if (strlen(first_name) >= sizeof(filename) - (2 + LEAF_LEN))
>> ++if (strlen(first_name) >= sizeof(filename) - (2 + LEAF_LEN)) {
>
> kind of curious that limit2 wasn't used here...

Good point, I reported this upstream:
https://lists.gnu.org/archive/html/bug-ncurses/2018-02/msg00016.html.

Cheers,
   Sven