714570 - Forward feed article not include the full text but the URL only

Status: RESOLVED FIXED

(MailNews Core :: Feed Reader, defect)

Description

[Lu Wei]

Attachment: the feed article I want to forward

User Agent: Mozilla/5.0 (Windows NT 5.1; rv:8.0) Gecko/20100101 Firefox/8.0
Build ID: 20111104165243

Steps to reproduce:

Forward a feed article that has full text content


Actual results:

The forward message only contains the article's URL


Expected results:

The forward message should contains the full article

Comment 1

[Lu Wei]

In TB support group, news: <[e-mail]> 
"alta88[nntp]" pointed out: 

"you have discovered an interesting bug.  if the messageId (View->Source 
or Show All Headers to see it) contains the url of the article (must be 
a valid url) then Forward includes it only.

if the messageId is anything else (invalid url or a 'tag:' format 
value), the full Summary of the article will filled in upon Forward.

the value used for messageId is based on what the publisher provides. 
but, for this purpose, Tb would need to do a workaround.  you should 
file a bug."

Comment 2

[Hashem Masoud]

(In reply to LU Wei from comment #1)
In my opinion, this should be an enhancement or an addon. It is not a bug.

Comment 3

[Lu Wei]

(In reply to Hashem Masoud from comment #2)
> (In reply to LU Wei from comment #1)
> In my opinion, this should be an enhancement or an addon. It is not a bug.

Whether you think the "enhancement" should make the forwards just forward a URL or full text, the present behavior (sometimes forward a URL, sometimes forward full text, according to "alta88[nntp]") is a bug.

Comment 4

[alta88]

this issue exists due to a Very Old bug 258278.  the conditions of that bug no longer exist, ie a summary is now always stored with the message to allow a toggle between summary and web page, since bug 438429.

the poor implementation in bug 258278 allowed forwarding summaries with tag: messageIds to work, but those feeds with http url messageIds got the link.  a feed messageId can be any construct.

if bug 258278 were reverted, then a forward would always include the summary (even if viewing a web page).  i can look into this at some point; forwarding a summary should definitely be possible.

Comment 5

[Lu Wei]

IMO, the default forward behavior should be the same as view setting, i.e. when in summary view, "forward" will forward the summary; when in web page view, "forward" will forward the web page.

Comment 6

[alta88]

Attachment: patch.

this patch sorts out how feed messages are determined for further compose handling.  it applies only to feed account messages (feeds moved/copied to other account types get normal handling, as inline messagepane web page viewing only happens in feed accounts).  if a user is viewing a summary, single/multiselection compose will load summary(ies); if viewing a web page, load link(s).

@comment 5: i too think there are nice aspects to loading the entire web page, but there are also issues, and don't think it would meet with approval by those who decide.  a full web page would be better done in an extension.  opening a link can be done in a content tab, which can allow javascript, and mitigates the usability problem somewhat.

Comment 7

[Lu Wei]

Put "web page view" situation aside (anyway, the feed message body contains only  "summary"), I hope the patch will apply to all account folders. Feed message does not differ from mail message in storage format, in user experience (at simple HTML view and plain text view, which I prefer), and filters are normally used to re-organize messages to different folders, whatever the account they are under.

Comment 8

[alta88]

as mentioned, a feed article moved to a non feed account is treated normally, ie as any other message in that account.

it is actually not strictly possible to identify a message as a 'feed' outside its account.  content-base in the header, of course, is a good indicator, but that header can be sent along by anyone.  see bug 522645.

Comment 9

[alta88]

Attachment: tweak to handle empty fwd prefix pref.

Comment 10

[alta88]

Attachment: better tweak.

Comment 11

[Lu Wei]

(In reply to alta88 from comment #8)
> as mentioned, a feed article moved to a non feed account is treated
> normally, ie as any other message in that account.
> 
Thanks, that's good.

> it is actually not strictly possible to identify a message as a 'feed'
> outside its account.  content-base in the header, of course, is a good
> indicator, but that header can be sent along by anyone.  see bug 522645.

The current situation is TB identifies this, because if a feed message is moved to a mail account, then forwarded, the summary body is also cut off and just URL will be forwarded.

Comment 12

[Ludovic Hirlimann [:Usul]]

(In reply to alta88 from comment #10)
> Created attachment 588760 [details] [diff] [review]
> better tweak.

Alta88 could we get some unit tests too ?

Comment 13

[alta88]

if someone could point me to where compose window open type tests are, i could take a look to incorporating this.

Comment 14

[Magnus Melin [:mkmelin]]

http://mxr.mozilla.org/comm-central/source/mail/test/mozmill/composition/ maybe?

Comment 17

[alta88]

Attachment: updated patch

updated patch to unbitrot and better handle empty content-base header.

Comment 18

[David :Bienvenu]

Comment on attachment 601822 [details] [diff] [review]
updated patch

r=me, but doesn't createInstance throws when it fails? That would make this check not useful:

+    if (!params || !composeFields)
+      return;

Comment 19

[alta88]

it looks like it.  those checks were there since the beginning.  do you prefer them removed?

Comment 20

[David :Bienvenu]

(In reply to alta88 from comment #19)
> it looks like it.  those checks were there since the beginning.  do you
> prefer them removed?

yes, thx, I think it's better to remove them just so no one copies the bad pattern.

Comment 21

[alta88]

Attachment: update for checkin.

Comment 22

[Ryan VanderMeulen [:RyanVM]]

http://hg.mozilla.org/comm-central/rev/b3925ea79358

For future patches, please use a checkin comment that summarizes the changes made rather than reiterating the bug title.