bug#33133: 26.1.50; zlib-decompress-region too rigid

[Noam Postavsky <npostavs@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 608 bytes --]

tags 33133 + patch
quit

Katsumi Yamaoka <yamaoka@jpl.org> writes:

> On Tue, 23 Oct 2018 20:26:59 -0400, Noam Postavsky wrote:
>> --- i/src/decompress.c
>> +++ w/src/decompress.c
>> @@ -206,7 +206,7 @@ DEFUN ("zlib-decompress-region", Fzlib_decompress_region,
>
>>    while (inflate_status == Z_OK);
>
>> -  if (inflate_status != Z_STREAM_END)
>> +  if (inflate_status != Z_STREAM_END && inflate_status != Z_BUF_ERROR)
>>      return unbind_to (count, Qnil);
>
>>    unwind_data.start = 0;
>
> I confirmed that it makes it work for the corrupted web site in
> question.  Thank you!

Here's a proper patch.


[-- Attachment #2: patch --]
[-- Type: text/plain, Size: 4856 bytes --]

From 430ebd936b0bc41bd3e33e171938161846597196 Mon Sep 17 00:00:00 2001
From: Noam Postavsky <npostavs@gmail.com>
Date: Sat, 27 Oct 2018 17:45:52 -0400
Subject: [PATCH v1] Allow partial decompression (Bug#33133)

* src/decompress.c (Fzlib_decompress_region): Add optional
ALLOW-PARTIAL parameter.
* lisp/url/url-http.el (url-handle-content-transfer-encoding): Use it.
* doc/lispref/text.texi (Decompression): Document it.
* etc/NEWS: Announce it.
---
 doc/lispref/text.texi | 10 ++++++----
 etc/NEWS              |  6 ++++++
 lisp/url/url-http.el  |  5 +++--
 src/decompress.c      | 22 +++++++++++++++++-----
 4 files changed, 32 insertions(+), 11 deletions(-)

diff --git a/doc/lispref/text.texi b/doc/lispref/text.texi
index 6c38d8eed0..e39ba6a192 100644
--- a/doc/lispref/text.texi
+++ b/doc/lispref/text.texi
@@ -4462,14 +4462,16 @@ Decompression
 available.
 @end defun
 
-@defun zlib-decompress-region start end
+@defun zlib-decompress-region start end &optional allow-partial
 This function decompresses the region between @var{start} and
 @var{end}, using built-in zlib decompression.  The region should
 contain data that were compressed with gzip or zlib.  On success, the
 function replaces the contents of the region with the decompressed
-data.  On failure, the function leaves the region unchanged and
-returns @code{nil}.  This function can be called only in unibyte
-buffers.
+data.  If @var{allow-partial} is @code{nil}, on failure, the function
+leaves the region unchanged and returns @code{nil}.  Otherwise, it
+returns the number of bytes that were not decompressed and replaces
+the region text by whatever data was successfully decompressed.  This
+function can be called only in unibyte buffers.
 @end defun
 
 
diff --git a/etc/NEWS b/etc/NEWS
index 3f86195695..395169253d 100644
--- a/etc/NEWS
+++ b/etc/NEWS
@@ -1159,6 +1159,12 @@ to mean that it is not known whether DST is in effect.
 'json-insert', 'json-parse-string', and 'json-parse-buffer'.  These
 are implemented in C using the Jansson library.
 
++++
+** 'zlib-decompress-region' can partially decompress corrupted data.
+If the new optional ALLOW-PARTIAL argument is passed, then the data
+that was decompressed successfully before failing will be inserted
+into the buffer.
+
 ** Mailcap
 
 ---
diff --git a/lisp/url/url-http.el b/lisp/url/url-http.el
index 6b5749e1bc..94ac660fcf 100644
--- a/lisp/url/url-http.el
+++ b/lisp/url/url-http.el
@@ -939,7 +939,8 @@ url-http-parse-headers
     (goto-char (point-min))
     success))
 
-(declare-function zlib-decompress-region "decompress.c" (start end))
+(declare-function zlib-decompress-region "decompress.c"
+                  (start end &optional allow-partial))
 
 (defun url-handle-content-transfer-encoding ()
   (let ((encoding (mail-fetch-field "content-encoding")))
@@ -951,7 +952,7 @@ url-handle-content-transfer-encoding
 	(widen)
 	(goto-char (point-min))
 	(when (search-forward "\n\n")
-	  (zlib-decompress-region (point) (point-max)))))))
+	  (zlib-decompress-region (point) (point-max) t))))))
 
 ;; Miscellaneous
 (defun url-http-activate-callback ()
diff --git a/src/decompress.c b/src/decompress.c
index 2836338216..3872014739 100644
--- a/src/decompress.c
+++ b/src/decompress.c
@@ -120,12 +120,18 @@ DEFUN ("zlib-available-p", Fzlib_available_p, Szlib_available_p, 0, 0, 0,
 
 DEFUN ("zlib-decompress-region", Fzlib_decompress_region,
        Szlib_decompress_region,
-       2, 2, 0,
+       2, 3, 0,
        doc: /* Decompress a gzip- or zlib-compressed region.
 Replace the text in the region by the decompressed data.
-On failure, return nil and leave the data in place.
+
+If optional parameter ALLOW-PARTIAL is nil or omitted, on failure,
+return nil and leave the data in place.  Otherwise, return the number
+of bytes that were not decompressed and replace the region text by
+whatever data was successfully decompressed.  If decompression is
+completely successful return t.
+
 This function can be called only in unibyte buffers.  */)
-  (Lisp_Object start, Lisp_Object end)
+  (Lisp_Object start, Lisp_Object end, Lisp_Object allow_partial)
 {
   ptrdiff_t istart, iend, pos_byte;
   z_stream stream;
@@ -206,8 +212,14 @@ DEFUN ("zlib-decompress-region", Fzlib_decompress_region,
     }
   while (inflate_status == Z_OK);
 
+  Lisp_Object ret = Qt;
   if (inflate_status != Z_STREAM_END)
-    return unbind_to (count, Qnil);
+    {
+      if (!NILP (allow_partial))
+        ret = make_int (iend - pos_byte);
+      else
+        return unbind_to (count, Qnil);
+    }
 
   unwind_data.start = 0;
 
@@ -218,7 +230,7 @@ DEFUN ("zlib-decompress-region", Fzlib_decompress_region,
   signal_after_change (istart, iend - istart, unwind_data.nbytes);
   update_compositions (istart, istart, CHECK_HEAD);
 
-  return unbind_to (count, Qt);
+  return unbind_to (count, ret);
 }
 
 \f
-- 
2.11.0