• last updated 3 hours ago
Constraints
Constraints: committers
 
Constraints: files
Constraints: dates
fix bug, where months higher equal 08 lead to problems due to octal interpretation (can't use invalid octal number as operand...), aborting immediate ma

il sends in aug. or later.

fix bug, where months higher equal 08 lead to problems due to octal interpretation (can't use invalid octal number as operand...), aborting immediate mail sends in aug. or later.

Fix 3087 reordering the insert by putting the clob ones (cc, bcc, body and extraheaders) at the end to avoid crash on oracle 10g.

    • -10
    • +10
    ./acs-mail-lite-procs-oracle.xql
Merged from 5.4.2

  1. … 280 more files in changeset.
Add feature to allow incoming email to create new bugs

  1. … 5 more files in changeset.
mime::encode (tclllib 1.8 and 1.10) doesn't encode the question mark which is a character used as a delimiter for the content of the subject of an email message. That would cause email clients to interpret it incorrectly.

Removed the check for tcllib 1.10 to use our proc in all cases.

Add reply_to column necessary to sweeper proc in call acs_mail_lite::send_immediately

Ooops, forgot to rename a variable

Reodered the steps to build the message parts first, then add the headers using mime::setheaders (except To, From, CC and BCC). Retrieve the message as a string only if delivery mode is not set to default (smtp). Tested successfully with tcllib 1.8 and 1.10 (unless I missed something)

After discussion in the OCT, rolled this back. And leave it rolled back.

Fixed problem with complex send not knowing about the new valid_email_p procedure. additionally made sure that no empty header could be passed, as this breaks smtp::sendmail

Fix bug in subject encoding proc

Added an explicit blank body to two callbacks added by Malte, so automated

tests will pass. This is about the third release in which I've had to do

this. Simply adding "-" after the empty doc comment block is sufficient,

ad_proc treats that as an empty body.

Removed these two procs entirely, since a search of packages and contrib shows

that nothing in our tree uses them (and no package outside of acs-mail-lite

should've used them in the first place).

Use DCC header instead of the BCC one in call to ::smtp::sendmessage

Move complex_smtp to smtp. Mark ::sendmail and ::deliver_mail as deprecated

Upgrade scripts and fix datamodel consistency

  1. … 4 more files in changeset.
Remove duplicated proc

Remove unused queries. Remove from generic xql queries that are different for each db and already in the appropriate xql

Remove sendmail procs to use smtp only. Remove SendmailBin parameter (not used anymore)

    • -0
    • +34
    ./apm-callback-procs.tcl
  1. … 1 more file in changeset.
Replace send procs with a refactored version of complex_send, allowing cc, bcc and attachments

    • -23
    • +52
    ./acs-mail-lite-callback-procs.tcl
    • -25
    • +47
    ./acs-mail-lite-procs-oracle.xql
    • -25
    • +48
    ./acs-mail-lite-procs-postgresql.xql
  1. … 1 more file in changeset.
Update calls to utils procs adding the namespace

Move helper procs for sending email to acs-mail-lite

    • -0
    • +157
    ./utils-procs.tcl
  1. … 1 more file in changeset.
Rolling back to the working version (decided during informal OCT in IRC today)

  1. … 1 more file in changeset.
Add missing join

Fix typo, llength spelled incorrectly.

Cleaned up complex_send, broke it down into multiple parts, refactored repetitive things into foreach loops, made sure rollout support is handled in complex_smtp, thopugh I have no idea how to do the filters there.... This is in need of testing and should only act as a commit for the OCT to decide if this is going in the right direction.

Fix message initialization. Was failing in case the message has no alternative and no attachament. In case the message has an alternative and no attachment, it was initialized twice.

I think I fixed the problem with multipart/mixed of e-mails with no attachments. This should be tested though. I am putting this out to see if I might have a mismatch in my train of thought.

Fixed the confusion about single_email_p in complex_send_immediately using the defaults which make most sense. Fixed documentation accordingly